i
v
Lights, Camera, Reaction!
The Influence of Interfacial Chemistry on Nanoparticle Photoreactivity
by
Jeffrey Michael Farner Budarz
Department of Civil and Environmental Engineering Duke University
Date:_______________________ Approved:
___________________________ Mark R. Wiesner, Supervisor
___________________________
Claudia K. Gunsch
___________________________ Jerome Rose
___________________________
Richard T. Di Giulio
Dissertation submitted in partial fulfillment of the requirements for the degree of Doctor
of Philosophy in the Department of Civil and Environmental Engineering in the Graduate School
of Duke University
2016
i
v
ABSTRACT
Lights, Camera, Reaction!
The Influence of Interfacial Chemistry on Nanoparticle Photoreactivity
by
Jeffrey Michael Farner Budarz
Department of Civil and Environmental Engineering Duke University
Date:_______________________ Approved:
___________________________ Mark R. Wiesner, Supervisor
___________________________
Claudia K. Gunsch
___________________________ Jerome Rose
___________________________
Richard T. Di Giulio
An abstract of a dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of
Civil and Environmental Engineering in the Graduate School of Duke University
2016
Copyright by Jeffrey Michael Farner Budarz
2016
iv
Abstract The ability of photocatalytic nanoparticles (NPs) to produce reactive oxygen
species (ROS) has inspired research into several new applications and technologies,
including water purification, contaminant remediation, and self-cleaning surface
coatings. As a result, NPs continue to be incorporated into a wide variety of increasingly
complex products. With the increased use of NPs and nano-enabled products and their
subsequent disposal, NPs will make their way into the environment. Currently, many
unanswered questions remain concerning how changes to the NP surface chemistry that
occur in natural waters will impact reactivity. This work seeks to investigate potential
influences on photoreactivity – specifically the impact of functionalization, the influence
of anions, and interactions with biological objects - so that ROS generation in natural
aquatic environments may be better understood.
To this aim, titanium dioxide nanoparticles (TiO2) and fullerene nanoparticles
(FNPs) were studied in terms of their reactive endpoints: ROS generation measured
through the use of fluorescent or spectroscopic probe compounds, virus and bacterial
inactivation, and contaminant degradation. Physical characterization of NPs included
light scattering, electron microscopy and electrophoretic mobility. These systematic
investigations into the effect of functionalization, sorption, and aggregation on NP
v
aggregate structure, size, and reactivity improve our understanding of trends that
impact nanoparticle reactivity.
Engineered functionalization of FNPs was shown to impact NP aggregation, ROS
generation, and viral affinity. Fullerene cage derivatization can lead to a greater affinity
for the aqueous phase, smaller mean aggregate size, and a more open aggregate
structure, favoring greater rates of ROS production. At the same time however, fullerene
derivatization also decreases the 1O2 quantum yield and may either increase or decrease
the affinity for a biological surface. These results suggest that the biological impact of
fullerenes will be influenced by changes in the type of surface functionalization and
extent of cage derivatization, potentially increasing the ROS generation rate and
facilitating closer association with biological targets.
Investigations into anion sorption onto the surface of TiO2 indicate that reactivity
will be strongly influenced by the waters they are introduced into. The type and
concentration of anion impacted both aggregate state and reactivity to varying degrees.
Specific interactions due to inner sphere ligand exchange with phosphate and carbonate
have been shown to stabilize NPs. As a result, waters containing chloride or nitrate may
have little impact on inherent reactivity but will reduce NP transport via aggregation,
while waters containing even low levels of phosphate and carbonate may decrease
“acute” reactivity but stabilize NPs such that their lifetime in the water column is
increased.
vi
Finally, ROS delivery in a multicomponent system was studied under the
paradigm of pesticide degradation. The presence of bacteria or chlorpyrifos in solution
significantly decreased bulk ROS measurements, with almost no �OH detected when
both were present. However, the presence of bacteria had no observable impact on the
rate of chlorpyrifos degradation, nor chlorpyrifos on bacterial inactivation. These results
imply that investigating reactivity in simplified systems may significantly over or
underestimate photocatalytic efficiency in realistic environments, depending on the
surface affinity of a given target.
This dissertation demonstrates that the reactivity of a system is largely
determined by NP surface chemistry. Altering the NP surface, either intentionally or
incidentally, produces significant changes in reactivity and aggregate characteristics.
Additionally, the photocatalytic impact of the ROS generated by a NP depends on the
characteristics of potential targets as well as on the characteristics of the NP itself. These
are complicating factors, and the myriad potential exposure conditions, endpoints, and
environmental systems to be considered for even a single NP highlight the need for
functional assays that employ environmentally relevant conditions if risk assessments
for engineered NPs are to be made in a timely fashion so as not to be outpaced by, or
impede, technological advances.
vii
Dedication
No work is produced in a vacuum (don’t believe the particle physicists), and so
this dissertation is dedicated to everyone who has contributed to this endeavor. Truly
without your collaboration, ideas, and support, in ways both large and small, I would
not find myself here.
viii
Contents
Abstract .......................................................................................................................................... iv
List of Tables ................................................................................................................................ xv
List of Figures ............................................................................................................................. xvi
Acknowledgements .................................................................................................................. xxii
1. Introduction ............................................................................................................................... 1
1.1 Motivation ...................................................................................................................... 1
1.2 Scope ............................................................................................................................... 2
1.3 Hypotheses and Objectives .......................................................................................... 4
2. Background ................................................................................................................................ 7
2.1 Photocatalysis .................................................................................................................... 8
2.1.1 Reactive Oxygen Species .......................................................................................... 10
2.1.1.1 ROS Detection ..................................................................................................... 12
2.1.1.2 ROS Standard Generators ................................................................................. 15
2.1.1.3 ROS Quenchers .................................................................................................. 16
2.2 Photoreactive Nanoparticles ......................................................................................... 16
2.2.1 Fullerene Nanoparticles ........................................................................................... 19
2.2.1.1 FNP Reactivity .................................................................................................... 20
2.2.1.2 FNP Physical Transformations ........................................................................ 22
2.2.2 Titanium Dioxide ....................................................................................................... 23
2.2.2.1 TiO2 Reactivity .................................................................................................... 23
ix
2.2.2.2 P25 TiO2 ............................................................................................................... 25
2.3 Particle Stability and Aggregation ............................................................................... 27
2.3.1 Particle Attachment ................................................................................................... 27
2.3.2 Rectilinear Transport and Particle Collisions ........................................................ 30
2.3.3 Fractal Dimension ..................................................................................................... 32
2.3.4 Changes to the Smoluchowski Equation Due to the Fractal Nature of Aggregates ........................................................................................................................... 33
2.3.5 Measuring Fractal Dimension via Light Scattering .............................................. 35
2.4 Effect of Aggregation on Reactivity ............................................................................. 38
2.5 NPs in Environmental Waters ...................................................................................... 40
2.6 Possible Risks Posed by Nanomaterials ...................................................................... 42
2.6.1 Need for Standardized Systems and Tests – Functional Assays ........................ 44
3. Methods .................................................................................................................................... 46
3.1 NP Suspension Preparation .......................................................................................... 46
3.1.1 FNP Suspensions ....................................................................................................... 46
3.1.2 TiO2 Suspensions ....................................................................................................... 47
3.2 UV Illumination .............................................................................................................. 48
3.3 ROS Measurements ........................................................................................................ 49
3.3.1 1O2 Measurements with SOSG ................................................................................. 49
3.3.1.1 Standard Curve Preparation ............................................................................ 50
3.3.2 �OH Measurements with TA ................................................................................... 52
3.3.2.1 Standard Curve Preparation ............................................................................ 54
x
3.3.3 O2�- Measurements with HE .................................................................................... 54
3.4 Additional Reactivity Endpoints .................................................................................. 55
3.4.1 Viral Inactivation ....................................................................................................... 55
3.4.2 Bacterial Inactivation ................................................................................................ 56
3.4.3 Inactivation Kinetics and ROS Dose ....................................................................... 57
3.4.4 CPF Degradation ....................................................................................................... 57
3.5 Aggregate Characterization .......................................................................................... 58
3.5.1 Transmission Electron Microscopy ......................................................................... 58
3.5.2 Aggregate Size by Laser Diffraction ....................................................................... 58
3.5.3 Particle Size by Dynamic Light Scattering ............................................................. 59
3.5.4 Fractal Dimension by Static Light Scattering ........................................................ 59
3.5.5 Electrophoretic Mobility and pH Titrations .......................................................... 60
4. Results and Discussion ........................................................................................................... 62
4.1 Bacteriophage Inactivation by UV-A Illuminated Fullerenes: Role of Nanoparticle-Virus Association and Biological Targets ................................................. 62
4.1.1 Introduction ................................................................................................................ 63
4.1.2 Materials and Methods ............................................................................................. 66
4.1.2.1 Cultures, Phage Purification and Enumeration ............................................. 66
4.1.2.2 Fullerene Suspensions ....................................................................................... 66
4.1.2.3 Irradiation and 1O2 Detection ............................................................................ 66
4.1.2.4 ATR-FTIR Spectroscopy .................................................................................... 67
4.1.2.5 OxyBlot Protein Oxidation Assay .................................................................... 68
xi
4.1.2.6 SDS-PAGE ........................................................................................................... 69
4.1.2.7 Dark-Field based Hyperspectral Imaging Microscopy ................................ 69
4.1.3 Results and Discussion ............................................................................................. 70
4.1.3.1 1O2 Generation by UV-A Sensitized Fullerenes ............................................. 70
4.1.3.2 1O2 Mediated MS2 Inactivation ........................................................................ 73
4.1.3.3 Alterations to Protein Secondary Structures .................................................. 74
4.1.3.4 Formation of Carbonyl Groups and Protein Conformational Changes ..... 76
4.1.3.5 Evidence for Capsid Rupture or Deformation ............................................... 78
4.1.3.6 Direct Qualitative Evidence for Differential Association of FNPs with MS2........................................................................................................................................... 78
4.1.3.7 1O2 Delivery and Proximity Implications for Chick-Watson Kinetics ........ 80
4.1.4 Conclusions ................................................................................................................ 84
4.2 Influence of Inorganic Anions on the Reactivity of TiO2 Nanoparticles ................ 86
4.2.1 Introduction ................................................................................................................ 87
4.2.2 Materials and Methods ............................................................................................. 90
4.2.2.1 Reagent Solutions ............................................................................................... 90
4.2.2.2 TiO2 Dispersion and Suspension Characterization ....................................... 91
4.2.2.3 TiO2 Nanoparticle Reactivity Measurement .................................................. 93
4.2.2.4 Photogenerated Holes ....................................................................................... 93
4.2.2.5 Hydroxyl Radicals ............................................................................................. 94
4.2.3 Results and Discussion ............................................................................................. 95
4.2.3.1 TiO2 Aggregate Characterization ..................................................................... 95
xii
4.2.3.2 Aggregation in Inorganic Salts ......................................................................... 96
4.2.3.3 Fractal Dimension .............................................................................................. 98
4.2.3.4 Aggregate Stability .......................................................................................... 100
4.2.3.5 Influence of ROS Detection Method .............................................................. 102
4.2.3.6 TiO2 Photoreactivity – Photogenerated Holes ............................................. 103
4.2.3.7 TiO2 Photoreactivity - Hydroxyl Radicals .................................................... 107
4.2.4 Conclusions .............................................................................................................. 110
4.3 Chlorpyrifos Degradation via Photoreactive TiO2 Nanoparticles: Assessing the Impact of a Multi-Component Degradation Scenario ................................................... 113
4.3.1 Introduction .............................................................................................................. 114
4.3.2 Materials and Methods ........................................................................................... 116
4.3.2.1 TiO2 Suspension and Characterization ......................................................... 116
4.3.2.2 UV Irradiation and Photoreactivity ............................................................... 117
4.3.2.3 Bacterial Cultures ............................................................................................. 118
4.3.2.4 Chlorpyrifos ...................................................................................................... 120
4.3.2.5 Sample Preparation and Analysis ................................................................. 120
4.3.2.6 CAKE software package ................................................................................. 122
4.3.3 Results ....................................................................................................................... 122
4.3.3.1 TiO2 Characterization and Reactivity ............................................................ 122
4.3.3.2 CPF Interactions with the TiO2 Surface ........................................................ 123
4.3.3.3 CPF UV/TiO2 Degradation .............................................................................. 124
4.3.3.4 Impact of UV/TiO2 on Bacteria ....................................................................... 126
xiii
4.3.3.5 Impact of CPF and Degradation Products on Bacteria ............................... 129
4.3.3.6 Impact of Bacteria and CPF on TiO2 Reactivity ........................................... 129
4.3.3.7 Impact of Bacteria on CPF Degradation ....................................................... 130
4.3.4 Discussion ................................................................................................................. 131
4.3.4.1 TiO2 Photocatalysis .......................................................................................... 131
4.3.4.2 CPF Adsorption ................................................................................................ 133
4.3.4.3 CPF Degradation and Oxidation Products .................................................. 134
4.3.4.4 Comparison of TiO2 Reactivity in the Presence and Absence of CPF ...... 135
4.3.4.5 Reactivity in the Presence of CPF and Bacteria ........................................... 136
4.3.5 Conclusions .............................................................................................................. 137
5. Conclusions ............................................................................................................................ 138
5.1 Bacteriophage Inactivation by UV-A Illuminated Fullerenes: Role of Nanoparticle-Virus Association and Biological Targets ............................................... 138
5.2 Influence of Inorganic Anions on the Reactivity of TiO2 Nanoparticles .............. 140
5.3 Chlorpyrifos Degradation via Photoreactive TiO2 Nanoparticles: Assessing the Impact of a Multi-Component Degradation Scenario ................................................... 142
5.4 Summary of Conclusions ............................................................................................ 143
5.5 Continuations of This Work ........................................................................................ 143
5.5.1 Natural Organic Matter .......................................................................................... 143
5.5.2 Natural Colloids ...................................................................................................... 144
5.5.3 Photoreactivity in Natural Systems – Towards Functional Assays ................. 144
5.5.3.1 Functional Assay Endpoint Considerations ................................................ 145
5.5.3.2 Environmental Release Considerations ........................................................ 146
xiv
5.5.3.3 Nanoinformatics, Data Visualization, and Comparison ............................ 148
Appendix A – Supporting Information for Chapter 4.1: Bacteriophage Inactivation by UV-A Illuminated Fullerenes: Role of Nanoparticle-Virus Association and Biological Targets ........................................................................................................................................ 150
Appendix B – Supporting information for Chapter 4.2: Influence of Inorganic Anions on the Reactivity of TiO2 Nanoparticles ...................................................................................... 167
Appendix C – Supporting Information for Chapter 4.3: Chlorpyrifos Degradation via Photoreactive TiO2 Nanoparticles: Assessing the Impact of a Multi-Component Degradation Scenario ............................................................................................................... 186
References ................................................................................................................................... 193
Biography ................................................................................................................................... 222
xv
List of Tables Table 1: Selected ROS standard generators, probes, and quenchers. .................................. 13
Table 2: Light scattering regimes of fractal aggregates. ........................................................ 36
Table 3: Estimated values of the equilibrium adsorption coefficient, KA, for iodide, carbonate and phosphate and estimated values of LH reaction constants for iodide oxidation, KR. R2 values are the coefficients of determination. .......................................... 106
Table 4: Rate constants of hydroxyl radicals and inorganic anions. .................................. 107
Table 5: Multiple reaction monitoring transitions used in the quantification of CPF and its degradation products. ......................................................................................................... 121
Table 6: Electrophoretic mobility and zeta potential of FNP suspensions and MS2 virus. ...................................................................................................................................................... 151
Table 7: MS2 inactivation rate constant during control experiments in UV-A. Average rate constant and standard deviation are shown here. ....................................................... 157
Table 8: MS2 viability during control experiments in the dark. ......................................... 157
Table 9: Infrared frequencies (wavenumbers) and associated band assignments for MS2 bacteriophage. ............................................................................................................................ 161
Table 10: Location of amino acids (AA) on A- and coat-protein sequences that have the potential to react with 1O2 at high second order reaction rates at physiological pH 7.1 [273, 373] ..................................................................................................................................... 163
Table 11: Initial values of CAKE sequence parameters ....................................................... 190
Table 12: Estimated values from CAKE software. ............................................................... 190
Table 13: χ ≤ the listed values from CAKE software. ........................................................... 190
Table 14: Decay time from CAKE software. .......................................................................... 191
Table 15: Additional statistics from CAKE software. .......................................................... 191
xvi
List of Figures Figure 1: Physical and chemical relaxation pathways of excited species. Adapted from [49]. .................................................................................................................................................. 9
Figure 2: Terephthalic acid hydroxylation produces the fluorescent 2-hydroxy-terepthalic acid. Adapted from [89]. ........................................................................................ 14
Figure 3: Reduction of XTT by superoxide. Adapted from [73]. .......................................... 14
Figure 4: Oxidation of SOSG by 1O2. Adapted from [83]. ...................................................... 15
Figure 5: (a) C60, fullerene and (b) C60(OH)24, fullerol. ........................................................... 20
Figure 6: Photosensitization of and ROS production by C60. Adapted from [81]. ............. 21
Figure 7: Photosensitization of and ROS production by TiO2. Adapted from [40]. .......... 25
Figure 8: �OH production of various TiO2 suspensions at 20ppm, measured by TA and graphed as fluorescence intensity vs. emission wavelength after 5 minutes of UV-A exposure. Unpublished data. .................................................................................................... 26
Figure 9: DLVO schematic illustrating the sum (solid black line) of electrostatic repulsion (blue dashed line) and van der Waal’s attraction (red dashed line) as a function of separation distance. .................................................................................................................... 28
Figure 10: TEM images of 25nm hematite NPs aggregated under (a) DLA (Df = 1.68) and (b) RLA (Df = 2.20) schemes. Adapted from [198]. ................................................................. 32
Figure 11: Schematic of rectilinear (far left) and curvilinear (far right) transport models alongside calculations of flow lines in fractal aggregates with radially varying permeability. Adapted from [210] and [209]. .......................................................................... 34
Figure 12: Illustration of light scattering regimes when plotting SLS data as intensity vs. scattering vector. Adapted from [196] ..................................................................................... 37
Figure 13: 1O2 generation rates in UV-A sensitized FNP suspensions (15 mM PBS) ........ 71
Figure 14: TEM images depicting the morphology of nanoparticle aggregates. Aqu-nC60 crystal lattice planes are visible in (e). ...................................................................................... 72
xvii
Figure 15: MS2 inactivation by UV-A alone and UV-A sensitized aqu-nC60, C60(OH)6, C60(OH)24, C60(NH2)6. ................................................................................................................... 74
Figure 16: Changes in protein secondary structures (intermolecular extended chains, β-sheets and random coils, α- and 3-turn helices, and antiparallel β-sheets and aggregated strands following 5 minute exposure to UV-A illuminated FNP suspensions from ATR-FTIR. CT values correspond to 10-9 mg min L-1. ..................................................................... 75
Figure 17: (a) Evolution of carbonyl groups in MS2 upon 1O2 oxidation from ATR-FTIR. (b) The extent of MS2 protein oxidation by 1O2 at the end of 5 min of UV-A sensitization is shown here in terms of carbonyl concentrations from the OxyBlot assay. CT values correspond to 10-9 mg min L-1. ................................................................................................... 76
Figure 18: ATR-FTIR spectra of MS2 exposed to UV-A illuminated FNPs showing the development of amino acid residues and RNA compared to UV-A alone. ....................... 78
Figure 19: Relative proximity of various fullerenes with MS2 following MTMF analysis of hyperspectral images. 5-fold zoom-ins (1 cm = 20 µm) of noise-free image data with superimposed classification results are shown. MS2 viruses appear as green pixels after staining with 2.5% phosphotungstic acid. (a) C60(NH2)6 is shown in red, (b) aqu-nC60 in blue, (c) C60(OH)6 in purple, and (d) C60(OH)24 in pink. (e) –(h) Corresponding TEM images after sample dehydration. ............................................................................................ 79
Figure 20: Evaluation of the Chick-Watson model to describe MS2 inactivation kinetics by all four nanoparticles. ........................................................................................................... 83
Figure 21: Conceptual interplay between fullerene derivatization, quantum yield, and 1O2 generation. ............................................................................................................................. 84
Figure 22: TiO2 aggregation over time for each anion at (a) 0.5 mM and (b) 25 mM. Data shown is volume weighted D50 (mean±std) ............................................................................ 97
Figure 23: TiO2 aggregation (volume weighted D50 mean±std) vs. time as a function of chloride concentration. ............................................................................................................... 98
Figure 24: Fractal dimension of TiO2 aggregates for all anions at 12.5 and 25 mM. Df values are the average of the final 3 measurements (mean±std). ........................................ 99
Figure 25: (a) Zeta potential vs. pH for all anions at 0.5 mM. (b) IEP vs. concentration for all anions. IEP for 0.5 mM phosphate was not determined as the ζ potential was negative over the range of pH tested. .................................................................................................... 101
xviii
Figure 26: Oxidized iodide concentration (mean±std) versus anion concentration after 30 minute irradiation (50 mM KI) as a function of species in solution (carbonate, phosphate). The blue dotted line is the iodide concentration of TiO2 suspensions containing only 50 mM KI. ...................................................................................................... 104
Figure 27: (a) 2-Hydroxyterephthalic acid concentration (mean±std) vs. anion concentration after 30 minute irradiation as a function of species in solution. (b) Hydroxyl radical generation normalized to 0.5 mM anion concentration (mean±std). . 109
Figure 28: Supernatant CPF concentration vs. time as a function of TiO2 concentration (0 - 40 mg L-1 dark and 20 mg L-1 UV). ........................................................................................ 124
Figure 29: CPF and degradation product concentration vs. time in suspensions of 20 mg L-1 TiO2 exposed to UV. ............................................................................................................ 125
Figure 30: Potential degradation pathways considered by CAKE software. ................... 126
Figure 31: Inactivation of bacteria (A. baumannii and B. subtilis) under the following conditions: (UV+) 20 mg L-1 exposed to UV, (SOD) 20 mg L-1 TiO2 exposed to UV in the presence of SOD, (N-AC) 20 mg L-1 TiO2 exposed to UV in the presence of N-AC, (D+) 20 mg L-1 TiO2 in the dark, and (UV-) 0 mg L-1 TiO2 exposed to UV. ..................................... 127
Figure 32: TEM images of B. subtilis with 20 mg L-1 TiO2 (a) prior to UV exposure and (b) after 60 minutes UV exposure. Scale bars are 500nm. ......................................................... 128
Figure 33: Reactivity measurements of 20 mg L-1 TiO2 exposed to UV. (a) Detected �OH in the presence of various constituents. (b) Detected O2�- and �OH in the presence of CPF over 60 minutes. ................................................................................................................ 130
Figure 34: Degradation of CPF in the presence and absence of bacteria. ......................... 131
Figure 35: (a) Number weighted intensity and (b) unweighted intensity of fullerenes in suspension measured using dynamic light scattering at 90 degrees scattering angle. ... 151
Figure 36: Change in fluorescence measured over one minute irradiance versus rose bengal standard concentration. Error bars indicate the standard deviation of triplicate measurements. ........................................................................................................................... 152
Figure 37: Rate of change in fluorescence measured over one-minute irradiance vs. calculated singlet oxygen production by the RB standards indicated in Figure 39 and
xix
calculated by Equations 34 and 35. Error bars indicate the standard deviation of triplicate measurements. .......................................................................................................... 153
Figure 38: Output spectrum of the Philips 15W TL-D BLB bulbs used in the study. Output is centered at 365nm, and UV-A irradiance was measured at 2.3mW/cm2. ....... 154
Figure 39: Change in fluorescence over given steady state singlet oxygen concentrations, determined from a Rose Bengal standard curve. Error bars indicate the standard deviation of triplicate measurements. The slope of this line is k’ as indicated in Equation 39. Error bars indicate the standard deviation of triplicate measurements. .................... 156
Figure 40: Hyperspectral image analysis of MS2+aqu-nC60 in aqueous phase. Mixture tuned matched filtering was applied to map the distribution of MS2 and aqu-nC60 in the image. .......................................................................................................................................... 160
Figure 41: ATR-FTIR spectra of MS2 alone and MS2+fullerenes after UV-A exposure for 5 minutes showing evidence for carbonyl content and protein secondary structures. .. 162
Figure 42: Analysis by SDS-PAGE showing the altered migrations of MS2 capsid proteins (A-protein and coat protein) following exposure to UV-A sensitized 1O2. Lane A: molecular mass markers; Lane B: MS2 in Dark; Lane C: MS2+C60(NH2)6 in dark; Lane D: MS2+UV-A; Lane E: MS2+ C60(NH2)6+UV-A; Lane F: MS2+C60(OH)6+UV-A; Lane G: MS2+aqu-nC60+UV-A; Lane H: MS2+ C60(NH2)6+β-carotene+UV-A; Lane I: MS2+ C60(NH2)6+UV-A ........................................................................................................................ 166
Figure 43: Radiation intensity vs. emission wavelength at the upper liquid surface of samples. ...................................................................................................................................... 167
Figure 44: TiO2 aggregate size as a function of KI concentration (0, 50 mM) at various times (0, 30 min) (a,b). TiO2 aggregate D50 vs. time as a function of KI concentration (0, 50 mM) (c, d). Experimental results are expressed both as number weighted (a, c) and volume weighted (b, d). ........................................................................................................... 168
Figure 45: TiO2 aggregate size over 30 minutes as a function of anion at 0.5mM, (a) number weighted D50, and (b) volume weighted D50. ......................................................... 169
Figure 46: TiO2 aggregate size over 30 minutes as a function of anion concentration, (a, c, e, g, i) number weighted D50 and (b, d, f, h, j) volume weighted D50. ................................ 171
Figure 47: TiO2 aggregate size over 14 minutes as a function of anion concentration, volume weighted D50 for (a) nitrate and (b) sulfate. ............................................................ 172
xx
Figure 48: Potential energy curves of DLVO theory using the Gouy Chapman approximation for 20 mg L-1 TiO2 as a function of KCl concentration (0.5 – 25 mM). (a) Full plot and (b) detail. Blue dashed lines are electrostatic repulsion, red dashed lines are Van der Waals attraction, and black solid lines are the summation of repulsive and attractive forces for a given ionic strength. ........................................................................... 174
Figure 49: Representative SLS data and calculated fits at 0 and 30 minutes for each anion. ...................................................................................................................................................... 178
Figure 50: Zeta potential vs. pH as a function of anion concentration .............................. 179
Figure 51: (a) Absorption from iodo-starch method vs. time as a function of initial iodide concentration and (b) photogenerated hole detection (measured as oxidized iodide concentration) after 30 minutes irradiation versus initial concentration. ......................... 180
Figure 52: Oxidized iodide concentration (mean±std) vs. time as a function of inorganic anion concentration: (a) carbonate and (b) phosphate. ....................................................... 180
Figure 53: Experimental data linear regressions (1/r vs 1/[I-]0) for iodide adsorption via Langmuir-Hinshelwood mechanism for sonicated TiO2. ................................................... 181
Figure 54: Experimental data linear regressions (1/r vs. [Anion]) for the competitive adsorption of inorganic anions with iodide via Langmuir Hinshelwood mechanism for (a) carbonate and (b) phosphate. ............................................................................................ 183
Figure 55: 2-Hydroxyterephthalic acid concentration (mean±std) vs. time as a function of inorganic anion concentration for various anionic species: (a) carbonate, (b) chloride, (c)nitrate, (d) sulfate, and (e) phosphate. ............................................................................... 184
Figure 56: Normalized hydroxyl radical generation vs. inorganic anion concentration after 30 minute irradiation time as a function of anionic species in solution. ................. 185
Figure 57: TiO2 Characterization: (a) time resolved DLS, (b) TEM image of TiO2 aggregate in MD media; scale bar is 50nm, (c) DLS of TiO2 in MD media, unweighted, (d) ZP vs. pH of 20 ppm TiO2 in MD, adjusted with HCl. .................................................. 186
Figure 58: Reactivity measurements of hydroxyl radical and superoxide from 20 ppm TiO2 in MD exposed to UV, exposed to UV in the presence of quenching agents, and in the dark. ...................................................................................................................................... 187
Figure 59: Number weighted TRDLS of 20 ppm TiO2 in the presence of 385 ppb CPF. 187
xxi
Figure 60: (a) CPF, CPF oxon, and TCP over time in the dark with 0 ppm TiO2 (grey icons) and 20 ppm TiO2 (black icons). (b) Relative working stock concentrations measured at 0, 60 minutes in the dark. .................................................................................. 188
Figure 61: Bacterial inactivation due to 20 ppm TiO2 exposed to UV in the presence and absence of CPF (a) B. subtilis and (b) A. baumanii. ................................................................ 189
Figure 62: CAKE software fit of CPF, CPF Oxon, and TCP concentration versus time. Calculated half-lives are presented in the inset. ................................................................... 189
Figure 63: Bacterial inactivation tests under dark conditions in the presence of CPF Oxon and TCP (a) B. subtilis and (b) A. baumanii. ........................................................................... 192
xxii
Acknowledgements
While there are numerous individuals to thank who contributed directly to the
science contained herein, in all likelihood there are an equal, if not greater, number who
set the stage and who are similarly culpable. This is by no means an exhaustive list, but
it is an attempt to show my appreciation to those who played so integral a role. Bear
with me.
First and foremost, to my advisor, Mark Wiesner. Thank you for your support
these *ahem* several years and for allowing me the space to explore and figure it all out.
The manner in which you embrace your job is inspiring, and the confidence you have in
your students is enabling. You’ve created an environment that fosters professional (and
at least once, personal) relationships that is remarkably collaborative and supportive.
Thank you for the belief that any semi-reasonable excuse to travel (conferences,
collaborations, general knowledge) is sufficient, and that, once there it’s worth an extra
week to see and experience the location. Thank you, in the same vein, for continually
inviting others to come and spend time at Duke. My research and my life have both been
greatly enriched because of this. I can’t imagine having joined another research group,
and I’m honored to be a part of this amalgam of musicians and artists, of non-joiners
xxiii
and mushroom hunters, of Blockus champions, bon-vivants, and science geeks.
Wieslandia is a magical realm, and we are fortunate to have stumbled into it.
Secondly, to my committee, who have improved this work immeasurably. To
Claudia Gunsch, who has been unfailingly willing to meet and discuss research and
roadblocks and to provide advice. Thank you for your support when things were
moving slowly. To Jerome Rose, your input has been invaluable, and the interest you
have shown in this work and its success is deeply appreciated. Thank you for going out
of your way to enrich the times I’ve visited the CEREGE and for accommodating my
doomed experiments in your valuable beam time. To Rich Di Giulio, thank you for your
support and research suggestions. The opportunities you’ve provided through the
Superfund program have helped to add context to my work, and it’s inspiring to see
someone who clearly enjoys their job as much as you do.
I would especially like to thank those who have most supported this work
through their collaborations. To Soryong Chae, who provided my first glimpse into the
abyss that would be NP photoreactivity. This is arguably all your fault, so thank you. To
Raju Badireddy, a secret lover of heavy metal and a great collaborator and friend. I have
enjoyed working with you immensely and hope to do so again soon. To David Jassby,
the force of nature who can sleep anywhere. Your ability to view a project from a
distance and to see the missing piece is inspiring. To Andrea Turolla, who seamlessly
balances science and adventure. I’ll be lucky to end up half the researcher you are.
xxiv
Thank you for showing me Milan; I hope I can repay the favor soon. To Ellen Cooper,
for always falling for my ruse of offering to buy you coffee when I really wanted to ask
for your help. To Courtney Gardner, the microbial expert, lover of books, and part time
weightlifter. Thank you for helping meow with all the bacteria work and for not killing
me every time I tried to change or redo an experiment.
Next, I need to thank Team Wiesner, aka, the Wiesnerds. You have provided
support, camaraderie, and frustrations befitting a family and have been, quite possibly,
the greatest part of this adventure. To Mathieu Therezien, who is both a modeler and a
model, the modern day Viking with whom graduate school will always be most closely
associated: from coursework and qualifying exams to aggregation theory. For someone
who pretends to dislike people so frequently, you, sir, have a heart of gold. I should
know, I’ve taken advantage of this countless times. Our ping-pong battles have become
legendary in my mind. Thank you for the fights over who pays for coffee and for the
conversations that followed. Your friendship is more appreciated than I can convey.
You’re still a giant nerd, though.
To Amalia Turner, who lives deliberately and collectively. Thank you for
demonstrating the courage to try something nobody around you knows anything about
and then being willing to talk about experiments gone horribly wrong (I know many of
them go right, but everyone loves to talk about those). More importantly, thank you for
the countless lunches, dinners, beers, ABP cookie breaks, and the resulting conversations
xxv
that have enriched even small moments. Thank you for being someone who will debate
the relative quality levels of pizza in Durham. I’m so glad you joined our office, but also
glad you got the stubby desk.
To Judy Winglee, employee of the month forever, a rogue joiner in the midst of
the misfit toys, she of the church of flying disc, spikeball, and afternoon naps. I truly
enjoyed the discussions, and friendship, that arose from our writing our prelims
concurrently – I never realized Brooklyn backed up into Connecticut so seamlessly.
To Charles de Lannoy, he of the ascot and an unapologetic believer in beauty.
I’m inspired by your ability to balance the demands of coursework, research, and life
with earnest, and at times inebriated, grace and wit, and by your willingness to question
all things. While our experimental rabbit holes may not have led anywhere, the act of
jumping down them was both entertaining and enlightening.
To Zak Hendren, I’m thankful our friendship has transcended the basement of
Hudson. I admire you greatly. Let’s get together soon and drink some scotch.
To Christine Hendren, whose cheerful optimism in the face of adversity is only
one of several indicators of her amazing strength. You’re better at this science thing than
I think you realize, and it’s awesome to have had a front row ticket to watch. Thank you
for your unyielding support and overwhelming faith in people.
To Niall O’Brien, who is proof that perpetual motion may actually be possible. If
our research paths should intertwine again, I will be quite pleased indeed. To Shihong
xxvi
Lin, the most thoughtful inhabitant of Wieslandia. You’re brilliant and we all know it.
Conversations with you could go anywhere and often carried off in the perfect direction.
To Benjamin Espinasse, the pink-trucked boxer who can always find a deal. I’m still not
sure how the lab runs without you. Thank you for group lunches and farmhouse
dinners. To Stella Marinakos, thank you for putting up with us all. That you haven’t
strangled someone yet only underscores your patience. While we’re all feeling our way
through a pitch-black maze, you’re frequently the one with a flashlight. To Nick Geitner,
the only one who ever took Tie Tuesday seriously. Great things are in your future, and
you deserve them all. To Jacelyn Rice, who, when faced with a number of great
opportunities, had the guts to chose moving halfway across the world. I’m so excited for
you. To Nathan Bossa, whose easy laugh and eagerness has been a great addition to the
circle. To Alexis Carpenter, who showed us all up with her immaculate lab notebook, to
say nothing of her science. To Fabienne Schwab, the swiss cornhole terrorist. To Marielle
DuToit, the most hipster membrane scientist of them all.
To our French family – Jean-Yves Bottero, Melanie Auffan, Jerome Rose, Jerome
Labille, Armand Masion, and everyone at the CEREGE. It’s rare to find a lab that goes so
far out of their way to make you feel at home and to help you in your endeavors. I’m
grateful for your hospitality and for the time spent in Aix. It’s always a joy when you
come to Durham – that you are all world-class researchers is an added bonus.
xxvii
To everyone in the department who keeps the clockwork running. Without you,
we’d all be helpless. To Eileen Kramer, the gatekeeper who doesn’t know slow. Thank
you for your support (and ever-appreciated tip offs to free food). One of the best parts of
the CSSC has been getting to know you better. To Dwina Martin, who can fix anything,
or knows someone who can. I still hold you directly responsible for the most recent time
I’ve dislocated my shoulder, but I’m sure I owe you a diet coke for something, so let’s
call it even. To Ruby Nell Carpenter, who from first email to defense scheduling has
been a guide post in a confusing wilderness. Your dedication to the students of this
department is greatly undervalued.
To Duke cycling, you’ve kept me sane, provided an outlet when I desperately
needed one, and have been the most overwhelmingly positive addition to my life here.
I’ll cherish the countless hours, either on a bike, in a van, or standing in the blazing sun
or freezing rain. To Zhan Khaw and Kevin McDonnell, from those first rides, to trips to
the mountains, to nationals in Richmond, to Pitbull. To Matt Rinehart, possibly the
single most supportive and encouraging teammate Duke cycling will ever know. To
Kaleb Naegeli, K-swiss, the cartographer. You’re everything I like about Indiana, and
I’m thankful to have ridden with you and to know you. To Hoël Wiesner, potentially the
single greatest person to have on a ride, and an equally great person off the bike. Your
worldview is generally impeccable, although raccoons over opossums any day. Wizard
cats for life. To Jacob Miller, the only person I know with a greater love for puns and
xxviii
quite possibly the heart of Duke cycling. I’m glad to be leaving when you are. To
Chelsea Richwine, for do-rags and dried mango. For hotel tea and cookies, music
playlists back from West Virginia, and almost constant exasperation. I’m pretty sure
there’s a spider on your leg. To Laurel Beth Wheeler, Spike, my doppelganger/spirit
animal - she of cocktail hour and obsolete phones - I’m still only 72% sure you’re not a
secret agent. To Gina Turrini, G-rex, Mustang, the only one who needs coffee more than
me. Don’t worry; I’m not going to call you Peanut. To Ben Turits, the supposed
ringleader, for bluegrass, fire-and-brimstone preachers, and classic rock over the radio in
a cargo van with missing side view mirrors. I’m pretty sure this is just a way to never
grow up, and I like it. To Steven, Ginny, Robin, and Josh. To Becky and Owen and Matt,
Brenna and Dave. To Brandon, Mulvi and the Bananas. To Emma, Peter, and Tiki the
race dog. To all our frenemies in the ACCC. I’ve never felt more a part of a team, and
that wouldn’t be the case without you all. Really, the racing is secondary.
To Carley Gwin, who has navigated life, graduate school, and motherhood with
a grace not often seen. You’re doing a good job. Thank you for calling me out when I’ve
needed it and for many long hours writing at Cocoa Dos. To Gretchen Gehrke, who
takes on everything and somehow pulls it off, for long conversations about the idea of
grad school, the state of universities, of things to come, and what it all means. And for
equally long bike rides. To Milton and Ginger: you are, each in your own way, who I
hope to become. Thank you for letting us borrow your lawnmower that first weekend
xxix
we moved to Durham, for dinner shortly thereafter, and for everything that followed. To
ThuNDer, for the friendships that form when people eat together and the chaos that
arises when more than 5 people try to use the same stove.
To my family for, among so many things, your faith and the independence you
fostered in me, even though it’s come back to haunt you. To my Dad, for modeling the
importance of a college degree and for being willing to drive 6 hours late on Christmas
Eve to a small Kentucky town when my car breaks down so I can make it home in time
for Christmas. To my Mom, for supporting my decisions, even when it means moving
farther away, and for always telling me I was the best thing you ever did. I hope I’ve
lived up to that. To Nancy, for modeling the belief that if you want to do something
different in life, you can and, what’s more, you should. To Nathan, for modeling what
pursuing your passion looks like. To Amy, for the eagerness and completeness with
which you dive into life. I’m pretty sure you’re doing it right. To my Grandparents, for
never outright questioning when I’ll get a real job, and for always being proud of me. To
Grandpa Britt, for always believing that more education is a worthy endeavor and for
never meeting a cup of coffee you didn’t like. I’ve used that line countless times these
past years.
To Miiam, I’ve never known a more ardent supporter of her family, even when it
was in direct opposition to her personal happiness. I’m glad to fall within that circle. To
Timo, for making grad school look appealing and, in a total lie, easy. And to the entire
xxx
Budarz Clan, for being outrageously generous with your time. Through your unending
willingness to help, you have often modeled the best of what a family can be.
To Casa FaBu, I couldn’t imagine a finer house. A source of infinite frustration
and work, but also of accomplishment, joy and comfort. There was nothing more
satisfying than to physically tear down a wall when I was up against one in the lab, and
while I’ll straight up punch the previous owner in the nose if I ever see him, all the
unanticipated projects transformed you into my first home. You’ve hosted family,
parties, holiday dinners, ThuNDer, and countless Airbnb guests – some who became
good friends in a short period of time, and some who became good stories. I’d say
you’re finally finished, but we all know that would be a lie, and it doesn’t really matter
anyways.
To Luna, the would-be super villain who only ever cared for two people. It’s a
good thing you didn’t have claws, Lune-dog, else you’d have been much less lovable.
You’re sorely missed. To Sir Banjo Fiddlesticks. I can’t imagine a kinder, more
handsome muttly, and I’m sure even those you’ve humped along the way feel the same.
Thank you for following me around constantly no matter how long it’s been.
And most of all to Sara Farner Budarz, whose unfounded belief that I could do
this is perhaps the single greatest reason I find myself here. Thank you for being the
impetus for our relocation to North Carolina and the only one who would be foolhardy
and courageous enough to buy a house with me in a city we’ve never really seen just as
xxxi
we’re about to lose most of our income. Through beautiful, messy years of a half
destructed house, the rigors of us both in grad school, and the complexities of life, you
have remained my most ardent supporter. If this became my home, it is because of you.
If the world has been a more exciting place to explore, it is because of you. I’ve long
suspected you’d be the better academic, and I’m proud to see you proving me right.
That you live fully and unapologetically, only adds to your myth. Continue to demand
the most out of life, you deserve it.
And lastly, to Durm in the fine state of North Carolina, the first place I’ve not
wanted to leave. Small enough to pedal home and big enough to find anything you
need. Thank you for great shows, amazing restaurants, beautiful sunsets blocked out by
trees, bike races, soccer, and amazing people to experience it all with. These
acknowledgements were written during a farewell tour of sorts: at Motorco and
Fullsteam, Cocoa, Surf Club and Ponysaurus. Thank you for being the backdrop to some
amazing years.
1
1. Introduction
1.1 Motivation
Nanoparticles are not new. Nature has long coexisted with and employed
environmentally generated nanoparticles for many significant geological processes (e.g.,
cloud nucleation, heavy metal and radionucleotide transport, phytoplankton growth
leading to atmospheric CO2 regulation) [1]. Environmentally generated NPs arise from
many natural phenomena, including volcanoes and fire (CNTs and fullerenes) and
mineral weathering, redox chemistry, or biological interactions (metallic NPs) [1-6].
Our ability to manipulate NPs and produce them on a large scale, however, is
new. Since NPs were first discovered, their study and potential uses have driven
demand and production. Primarily, the appeal of NPs, and the incentive for their
adoption in technology and manufacturing, lies with their useful properties, such as
increased strength or reactivity [7]. Given the largely proprietary nature of NP
manufacturing, precise volumes are difficult to identify though estimates of total yearly
NP production are on the order of tens to hundreds of thousands of tons, corresponding
to hundreds of billions of dollars [7-9]. These NPs are incorporated into a wide variety of
increasingly complex products, though there is oftentimes considerable difficulty in
identifying which products use NPs and significant uncertainty in the composition of
the NPs included within consumer products [10-12].
2
1.2 Scope
Nanoparticle reactivity can refer to a variety of phenomena, such as electron
transfer from nanoparticulate zero valent iron (NZVI), catalysis by gold nanoparticles,
or photocatalysis by semiconducting NPs. An intriguing constraint specific to
photocatalysis is that it requires an external source of light to produce ROS - highly
reactive and unstable oxygen containing compounds that eagerly participate in
oxidation-reduction reactions. As a result of their instability, ROS have extremely short
lifetimes, often on the order of micro to milliseconds. The ability of some NPs to produce
ROS has inspired research aimed at using nanotechnology to inactivate viruses and
bacteria and to degrade contaminants in aquatic environments [13-16].
Whether by intentional design or through incidental interactions, nanoparticles
will not remain pristine for any appreciable length of time. NP functionalization,
coatings, and incorporation into matrices such as polymers or thin films will depend on
the use and desired functionality of user end products. All of these alterations will affect
the reactivity of the NP. The use and ageing of NP containing products will result in the
release of particles into the environment where further transformations are likely to
occur [9].
Risk is commonly viewed as a function of two contributing factors: hazard and
exposure [17-21]. For photocatalytic nanoparticles, transformations are expected to affect
both factors. For example, aqueous dispersions of fullerenes exposed to UV light have
3
been shown to undergo hydroxylation [22-24], which can lead to greater ROS
production (hazard). Alternatively, sorption of natural organic matter (NOM) has been
shown to affect the transport and reactivity of nanoparticles [25-28], affecting potential
levels of exposure. As a result, it is important to understand how the reactivity of a
nanoparticle is affected by its environment. Despite the investigations that have been
undertaken regarding NP reactivity, this area presents many challenges, and many
unanswered questions remain [29].
The primary motivation for this work is to investigate the influences on
photoreactivity (e.g. impact of functionalization, influence of anions, interactions with
biological objects) so that ROS generation in natural aquatic environments may be better
understood. Much of the research to date has proceeded on a case-by-case basis, focused
on modifying elemental makeup (doping) or functional groups to optimize ROS, or
looking at specific conditions to determine reactivity and ecotoxicity [30-35]. In this
work, the attempt has been to understand trends that impact nanoparticle reactivity. A
comprehensive risk assessment of NP photoreactivity falls outside the scope of this
work. Instead, photoreactivity is viewed as one aspect of any risk assessment involving
NPs that may undergo photo-excitation. As such, the primary focus of this work is on
the potential for photoreactivity to be used as a marker of eco-toxicity.
This research probes how surface functionalization, physicochemical sorption,
and aggregation of nanoparticles in the environment impact reactivity through
4
differences in characteristics such as aggregate size, aggregate morphology, and photo-
excitation. To achieve this, photocatalytic NPs - titanium dioxide (TiO2) and fullerene
nanoparticles (FNPs) - were studied in terms of their reactive endpoints: ROS
generation, virus and bacterial inactivation, and contaminant degradation.
Investigating these interactions has allowed for greater precision in determining the
reactivity of nanoparticles when considering their fate and transport in the environment.
Results from these experiments have improved our understanding of reactivity and
represent a step towards enabling ROS generation to be employed as a simple,
measurable endpoint for use in a functional assay approach to NP risk assessment. End
users of this research range from those who design and manufacture nanoënabled
technologies to aquatic ecologists and government entities studying water quality to
regulatory commissions who are currently grappling with establishing use and disposal
protocols.
1.3 Hypotheses and Objectives
The primary hypotheses of this work build from the understanding that surface
interactions will determine the reactivity of NPs in a system. Specifically, 1) that
functionalization of FNPs will reduce aggregation and impact ROS generation, 2) that
water chemistry and interactions with organic or ionic compounds at the NP surface will
determine the reactivity of the system, and 3) that multiple, coexisting constituents will
compete for ROS delivery.
5
ROS generation will be impacted by the size and structure of the aggregates that
are formed resulting from these interactions, in addition to any specific interactions that
alter the quantum yield and surface chemistry of the nanoparticles. These studies have
systematically investigated the impacts of functionalization, sorption, and aggregation
on NP aggregate structure, size, and reactivity - monitored as ROS generation as well as
compound degradation or microbiological inactivation. In doing so, specific interactions
have been identified and separated from aggregation affects. Ultimately, investigating
how these surface interactions affect reactivity furthers our understanding of the impact
from NP release into the environment.
To determine how engineered functionalization impacts NP aggregation and
ROS generation (hypothesis 1), stable suspensions of derivatized FNPs (in which various
function groups are attached to the C60 cage, i.e. aminated or hydroxylated fullerenes)
have been created. These suspensions were then characterized in terms of aggregate
size, electrophoretic mobility, ROS generation, and virus inactivation.
Investigations on how the sorption of compounds to the surface of a NP impacts
aggregation and reactivity (hypothesis 2) have been undertaken by monitoring TiO2
aggregation in various inorganic ions common to surface waters. Photoreactivity of TiO2
aggregates in the presence of the same ions was then measured both at the NP surface
and in the bulk solution and results were analyzed to determine the importance of
specific interactions between the NP surface and ion.
6
Finally, ROS delivery in a multicomponent system (hypothesis 3) was studied
under the paradigm of pesticide degradation. The impact of bacteria and chlorpyrifos on
bulk NP photoreactivity was determined individually before investigating the
implications of their combination with TiO2.
7
2. Background Photocatalysis is essentially the process of turning the energy in light into
chemical reactions. The ability of some materials to absorb and transfer the energy of a
photon has been known since the early 1900s, but the realization of the commercial
utility of photocatalysis is generally agreed to have origins in the studies using TiO2 to
split water by Fujishima and Honda in the early 1970s [36-38]. Many reviews exist that
cover the photocatalysis literature of FNPs and semiconductors [15, 16, 39-43]. Creating
and manipulating material at the nanoscale has resulted in both improved
electrochemical efficiency and increased interactions with contaminants in either air or
water, leading to applications that range from energy capture and storage, to water
treatment and contaminant degradation, to self-cleaning surfaces [38, 44-48]. When
considering NP photocatalysis, efficiency will depend on particle surface chemistry and
on the dynamics of particle interactions – particularly their tendency to form aggregates.
As a result, investigations should be designed so that potential influences are considered
and their impact on NP aggregation is accounted for.
This chapter provides an overview of nanoparticle photoreactivity and describes
the fundamental interactions that are expected to impact the reactivity of NPs in
aqueous environments. First, photocatalysis and the production of reactive oxygen
species (ROS) are described, including ROS detection and quantification. Secondly, the
two primary types of photoreactive nanoparticles used in this work, fullerene
8
nanoparticles and titanium dioxide, are introduced, each with an overview of their
photoreactivity and additionally pertinent physicochemical properties. Thirdly, the
fundamental processes leading to NP aggregation are described, followed by a
discussion of how aggregation and interfacial chemistry are expected to impact NP
photoreactivity. Lastly, efforts to understand the impact of NP photoreactivity in
complex, natural environments are discussed in light of the challenges associated with
NP risk assessment.
2.1 Photocatalysis
The Plank-Einstein equation describes the energy, E, of light with a given
wavelength, λ:
E = hcλ
(Eq 1.)
where h is plank’s constant and c, the speed of light. This energy is transferred to a
molecule through the absorbance of photons, exciting the electrons, which can follow a
variety of pathways to return to the ground state (Figure 1). This may take the form of
heat dissipation through vibration, light emission (fluorescence or phosphorescence), or
further energy transfer (sensitization) – all of which result in no net change to the
molecule. Alternatively, a molecule absorbing photons of sufficient wavelength may
undergo a chemical reaction resulting in fragmentation, rearrangement, or electron
transfer [49].
9
Figure+2.1+
Compound+ hν+(excita;on)+
Compound*+
Physical+
Processes+Chemical+
Processes+
Compound+ Products+
• Fragmenta;on+
• Hydrogen+abstrac;on+
• Electron+transDer+
• Rearrangement+
• Isomeriza;on+
• Dimeriza;on+
• Vibra;on+
• Light+Emission+
• Energy+Transfer+
Adapted+from+Schwarzenbach+(Ch15,+fig6)+
Figure 1: Physical and chemical relaxation pathways of excited species. Adapted from [49].
The absorption of light of sufficient energies has the capacity to degrade plastics
or pollutants and inactivate bacteria via direct photolysis, though the kinetics of such
reactions are often slow from an engineering standpoint [50-55]. However, reaction rates
may be increased, or unfavorable reactions facilitated, by the presence of catalysts in
suspension [49]. If the catalyst is of the same phase as the reactant, such as dissolved
iron participating in hydroxyl radical generation in the photo-Fenton reaction [56, 57],
the reaction is classified as homogeneous photocatalysis. Conversely, the reaction is
termed heterogeneous photocatalysis if the catalyst is of a different phase such as
photosensitized nanoparticles [40, 58]. This work focuses on the latter of these options
and how interactions at the surface of suspended NPs affect the photosensitized
reaction.
10
2.1.1 Reactive Oxygen Species
Heterogeneous photocatalysis in oxygenated waters primarily leads to the
production of reactive oxygen species. This can occur either by the sensitizer facilitating
electron transfer, denoted type I, or through the transfer of energy from the
photocatalyst, denoted type II [16, 59]. These two pathways will produce differing types
of ROS, but the net result in both cases is to convert absorbed light into chemical energy.
As a result, an effective photocatalyst will have both high light absorbance and high
quantum yields [49].
ROS is a family of oxygen containing compounds with unpaired or excited
electrons. ROS and reactive nitrogen species (RNS) play important roles in cellular
signaling and other biological processes, though they are also associated with cellular
toxicity [60-64]. On a larger scale, ROS is common in surface waters, created primarily
by the chromaphoric components of dissolved organic matter (CDOM), which have
been associated with maintaining biological balance and contaminant degradation. [65-
69].
The presence of excited or unpaired electrons in ROS destabilizes the compound
and leads to large redox potentials. As a result, radicals are short-lived, highly unstable,
and fairly unspecific. They readily react with other compounds as well as with water
itself, and the potential pathways of radical reactions are often complex [16, 70, 71]. The
11
three primary reactive oxygen species considered herein are singlet oxygen (1O2),
superoxide (O2�-), and hydroxyl radical (�OH).
Singlet oxygen is the result of type II sensitization in which an electron in the
oxygen molecule is excited from the π to the π* antibonding orbital [49]. This excitation
results in 1O2 being nearly 1V more oxidizing than ground state triplet oxygen [16, 72].
1O2 reacts most readily with carbon-carbon double and triple bonds, phenols, sulfides,
amines, and other anions [72]. 1O2 sensitizers are often organic dyes, aromatic
hydrocarbons and organic matter, though porphyrins and some inorganics such as
transition metal complexes and semiconducting metal oxides are also effective
generators [72].
O2 +P*→ 1O2 +P (Eq 2.)
Superoxide is the result of the single electron addition to molecular oxygen.
Sensitized dyes are readily capable of forming O2�- in the presence of an electron donor
[16]. A moderate reductant and nucleophile, O2�- may react through proton removal,
halide replacement, or as a single electron donor [59].
O2 + e− →O2
•− (Eq 3.)
The hydroxyl radical, the single electron oxidation product of hydroxide, is the
most oxidizing of the three primary forms of ROS (EH = 1.8V at pH 7) [16]. �OH reacts
primarily through the abstraction of a hydrogen atom, OH addition, or the one electron
oxidation to return to the hydroxide anion.
12
−OH + h+ → •OH (Eq 4.)
2.1.1.1 ROS Detection
ROS measurements, given the short half-lives involved, are often achieved via
probe compounds that are either oxidized or reduced upon reaction with the produced
radicals. These measurements are feasible because of the high reaction rate constants of
the probe compounds and the changes that result in spectroscopic properties (monitored
as fluorescence or absorption). Probe response generally indicates cumulative ROS
production over a measured exposure time, and care should be taken to limit the factors
(e.g. NP concentration, duration of exposure) that would result in an overabundance of
ROS production and hinder the probe’s reaction efficiency. Many probes are available
and their usage depends on experimental constraints such as bacterial toxicity, matrix
effects, pH, and solution chemistry [72-77]. Often these have been designed to monitor
endogenous ROS production for biological endpoints but can be used for exogenous
measurements. Other techniques, such as laser flash photolysis and electron
paramagnetic resonance spectroscopy (EPR) are effective but require extensive
instrumentation that makes them difficult to use with complex systems or in real time
environmental studies [78-80]. Monitoring specific compound degradation has been
effectively used to monitor multiple ROS species simultaneously, however liquid
chromatography is necessary for analysis, which can be time consuming [77, 81, 82].
13
Table 1: Selected ROS standard generators, probes, and quenchers.
Use Species Compound Reference
Probes 1O2 Singlet Oxygen Sensor Green [83, 84]
O2�- Tetrazolium salt, XTT [73, 85]
Hydrethidium [75, 86]
�OH Terephthalic Acid [76, 87-89]
Quenchers 1O2 β-Carotene [90]
L-Histidine [91, 92]
O2�- SOD [16, 81]
�OH N-Acetyl-l-cysteine [93]
Sodium Azide [94]
Generators 1O2 Rose Bengal [95-97]
Methylene Blue [72, 97]
O2�- Xanthine / Xanthine Oxidase [98]
�OH Hydrogen Peroxide [99, 100]
For measurements of hydroxyl radical production, the oxidation of terephthalic
acid (TA), shown in Figure 2, to 2-hydroxy terephthalic acid (2HTA), can be followed by
the fluorescence (315ex/425em) of the product. Using a standard curve, this fluorescence
can be quantified and tracked over time as a measure of cumulative hydroxyl radical
production [76, 87-89].
14
TA+Hydroxyla;on+
Figure+4.1+Oxida;on+of+TA+by+the+hydroxyl+radical+creates+the+fluorescent+2HTA+(315ex/425em).+Adapted+from+Eremia+et+al.38++
Figure 2: Terephthalic acid hydroxylation produces the fluorescent 2-hydroxy-terepthalic acid. Adapted from [89].
Superoxide production can be measured in a similar fashion using 2,3-bis(2-
Methoxy-4-nitro-5-sulphophenyl)-5-[(phenylamino)carbonyl]-2H-tetrazolium hydroxide
(XTT). Reduction of the compound by two superoxide molecules produces XTT-
formazon, which can be tracked via the formation of an absorption peak at 470nm
(Figure 3) [73, 85]. Alternatively, the reaction of superoxide with hydroethidine (HE) has
been shown to result in strong fluoresce (480ex/567em) [75, 86].
XTT+Reduc;on+/+ABS+
N
H
C
O
NN
NN
SO3-
SO3-
NO2
NO2
H3CO
H3CO
+
.
SO3
NO2
NO2
H3CO
H3CO
-
SO3-
O
C
H
N
N
N
N
N
2 O2 -.2 H+
2 O
+
2 H+ 2 O2.
Reduc;on+of+XTT+by+superoxide.+Adapted+from+Bartosz2006+
Figure 3: Reduction of XTT by superoxide. Adapted from [73].
Singlet oxygen generation can be measured using the florescent probe Singlet
Oxygen Sensor Green (SOSG) (Life Sciences, NY, USA). Oxidation of the weakly
fluorescent starting compound produces an endoperoxide (SOSG-EP), which fluoresces
15
much more strongly (Figure 4) [83, 95, 101, 102]. Though there are some concerns about
autoflourescence and indications that the molecule itself may act as a photosensitizer,
for lower concentrations of 1O2 and shorter timeframes, the response is linear and
efficient [83, 84].
SOSG+Oxida;on+
Figure+4.2+Oxida;on+of+SOSG+by+1O2+produces+the+fluorescent+endoperoxide+(ex504/em525).+Adapted+from+Gollmer+et+al.40++
moiety coupled to a light-emitting chromophore. Prior to thereaction with singlet oxygen, emission from the chromophoreis quenched by electron transfer from the adjacent trappingmoiety. Upon reaction with singlet oxygen, however, theresultant oxygen adduct is no longer an efficient intramolec-ular electron donor, and light emission readily occurs.
Two clear advantages of this indirect method to detectsinglet oxygen are (1) the luminescence quantum efficiency ofthe optical probe is comparatively large and (2) emissionoccurs in the visible region of the spectrum where opticaldetectors are very efficient. In turn, excellent optical signalscan be obtained from correspondingly smaller amounts ofsinglet oxygen in a given system. Of course, a limitation ofthis approach is that the probe may also respond to theeffects of reactive oxygen species other than singlet oxygen.Thus, a key aspect of characterizing such systems is toascertain the selectivity of the given probe to singlet oxygen(13–15).
The basic principle and structure of the two-componenttrap-fluorophore system (13) have since been exploited for thedevelopment of a commercially available product: SingletOxygen Sensor Green! (SOSG), Fig. 1 (15). Although thecompany that produces and sells SOSG has not explicitlyassociated a chemical structure with the name SOSG, themolecule shown in Fig. 1 is shown in the company’s patent(16) and, as outlined below, we have independently ascertainedthat this structure is indeed correct.
Although SOSG is a relatively new product, it has alreadybeen used in a number of published studies (17–21). Ofparticular importance, certainly in light of the work presentlydiscussed herein, is the report that SOSG can itself sensitize theproduction of singlet oxygen (22). Although the SOSG-sensi-tized singlet oxygen yields reported in this latter study are small(e.g. FD = 0.006 ± 0.002 upon 355 nm irradiation in aD2O ⁄CH3OH mixture), they are nevertheless not negligible.
In the present study, we further elaborate aspects of SOSGphotophysics and photochemistry, including features of singletoxygen production sensitized by (1) SOSG and (2) theimmediate product of the reaction between SOSG and singletoxygen, the SOSG endoperoxide (denoted SOSG-EP).
Although it has been reported that the behavior and responseof SOSG depends on the wavelength of irradiation (i.e. 355 vs532 nm) (22), we have generally limited our present study to anirradiation wavelength of 420 nm. This specific wavelengthwas chosen, in part, because porphyrins, and many porphyrinderivatives, are often used as singlet oxygen sensitizers, andthese molecules invariably have a pronounced absorption bandin this spectral domain (i.e. the Soret band). Moreover, manyendogenous singlet oxygen sensitizers have an appreciableabsorption cross section at 420 nm (23).
We also present data to indicate that SOSG can beincorporated into a living mammalian cell. This is an obser-vation that, heretofore, has not been reported; indeed, SOSG ismarketed as a cell-impermeable compound (15).
MATERIALS AND METHODSSample preparation. The SOSG used for all experiments reportedherein was purchased from Molecular Probes ⁄ Invitrogen. The purityand exact structure of this SOSG is unknown. In particular, in contrastto what is shown in Fig. 1 (i.e. 2,4-dicarboxylic acid on the phenylgroup), which is based on a structure shown in the Molecular Probespatent (16), it is likely that the material received consists of a mixtureof dicarboxylic acid isomers (see Supporting Information). Thematerial obtained was stored in a desiccator at £)20"C and protectedfrom light. The SOSG solutions used for the experiments wereprepared by first dissolving the compound in methanol (>99%; SigmaAldrich) and then diluting with D2O (>99%; Eurisotop) to yield a97:3 volume-to-volume D2O:CH3OH ratio. The residual alcohol isneeded to ensure complete dissolution of SOSG. Solutions of SOSGwere prepared under dim light (aerated solutions of SOSG arenevertheless reasonably stable upon exposure to ambient light).
Independently, SOSG was synthesized using the general protocolset forth in the patent taken by Molecular Probes (16). The details ofour approach are provided in the Supporting Information. For thepresent study, we only used this independently synthesized material toestablish the structure of SOSG.
5,10,15,20-Tetrakis(N-methyl-4-pyridyl)-21H, 23H-porphine(TMPyP, 98%) was purchased from Porphyrin Systems and used asthe independently added photosensitizer in the cell studies. Thephotosensitizer Phenalen-1-one-2-sulfonic acid (PNS), used as astandard for singlet oxygen quantum yield measurements, wassynthesized following a protocol described by Nonell et al. (24).Fluorescein (95%; Sigma Aldrich), bovine serum albumin (BSA, 99%;Sigma Aldrich) and NaN3 (>99%; Merck) were used as received.
ET
OHO
Cl
COOH
HOOC
O OHO O
Cl
COOH
HOOCO
O
weakly ylhgihtnecseroulf fluorescent
O2(a1 g)
ET
SOSG SOSG-EPFigure 1. Formation of the endoperoxide of SOSG (denoted SOSG-EP) upon reaction of SOSG with singlet oxygen. Prior to the reaction withsinglet oxygen, intramolecular electron transfer (ET) quenches the fluorescence from the light-emitting chromophore. Upon reaction with singletoxygen and the formation of the endoperoxide, electron transfer is precluded, and fluorescence is observed. As indicated in Materials and Methods,the SOSG commercially obtained is likely a mixture of the 2,4 and 2,5 dicarboxylic acids, not just the 2,4 isomer shown here.
2 Anita Gollmer et al.
moiety coupled to a light-emitting chromophore. Prior to thereaction with singlet oxygen, emission from the chromophoreis quenched by electron transfer from the adjacent trappingmoiety. Upon reaction with singlet oxygen, however, theresultant oxygen adduct is no longer an efficient intramolec-ular electron donor, and light emission readily occurs.
Two clear advantages of this indirect method to detectsinglet oxygen are (1) the luminescence quantum efficiency ofthe optical probe is comparatively large and (2) emissionoccurs in the visible region of the spectrum where opticaldetectors are very efficient. In turn, excellent optical signalscan be obtained from correspondingly smaller amounts ofsinglet oxygen in a given system. Of course, a limitation ofthis approach is that the probe may also respond to theeffects of reactive oxygen species other than singlet oxygen.Thus, a key aspect of characterizing such systems is toascertain the selectivity of the given probe to singlet oxygen(13–15).
The basic principle and structure of the two-componenttrap-fluorophore system (13) have since been exploited for thedevelopment of a commercially available product: SingletOxygen Sensor Green! (SOSG), Fig. 1 (15). Although thecompany that produces and sells SOSG has not explicitlyassociated a chemical structure with the name SOSG, themolecule shown in Fig. 1 is shown in the company’s patent(16) and, as outlined below, we have independently ascertainedthat this structure is indeed correct.
Although SOSG is a relatively new product, it has alreadybeen used in a number of published studies (17–21). Ofparticular importance, certainly in light of the work presentlydiscussed herein, is the report that SOSG can itself sensitize theproduction of singlet oxygen (22). Although the SOSG-sensi-tized singlet oxygen yields reported in this latter study are small(e.g. FD = 0.006 ± 0.002 upon 355 nm irradiation in aD2O ⁄CH3OH mixture), they are nevertheless not negligible.
In the present study, we further elaborate aspects of SOSGphotophysics and photochemistry, including features of singletoxygen production sensitized by (1) SOSG and (2) theimmediate product of the reaction between SOSG and singletoxygen, the SOSG endoperoxide (denoted SOSG-EP).
Although it has been reported that the behavior and responseof SOSG depends on the wavelength of irradiation (i.e. 355 vs532 nm) (22), we have generally limited our present study to anirradiation wavelength of 420 nm. This specific wavelengthwas chosen, in part, because porphyrins, and many porphyrinderivatives, are often used as singlet oxygen sensitizers, andthese molecules invariably have a pronounced absorption bandin this spectral domain (i.e. the Soret band). Moreover, manyendogenous singlet oxygen sensitizers have an appreciableabsorption cross section at 420 nm (23).
We also present data to indicate that SOSG can beincorporated into a living mammalian cell. This is an obser-vation that, heretofore, has not been reported; indeed, SOSG ismarketed as a cell-impermeable compound (15).
MATERIALS AND METHODSSample preparation. The SOSG used for all experiments reportedherein was purchased from Molecular Probes ⁄ Invitrogen. The purityand exact structure of this SOSG is unknown. In particular, in contrastto what is shown in Fig. 1 (i.e. 2,4-dicarboxylic acid on the phenylgroup), which is based on a structure shown in the Molecular Probespatent (16), it is likely that the material received consists of a mixtureof dicarboxylic acid isomers (see Supporting Information). Thematerial obtained was stored in a desiccator at £)20"C and protectedfrom light. The SOSG solutions used for the experiments wereprepared by first dissolving the compound in methanol (>99%; SigmaAldrich) and then diluting with D2O (>99%; Eurisotop) to yield a97:3 volume-to-volume D2O:CH3OH ratio. The residual alcohol isneeded to ensure complete dissolution of SOSG. Solutions of SOSGwere prepared under dim light (aerated solutions of SOSG arenevertheless reasonably stable upon exposure to ambient light).
Independently, SOSG was synthesized using the general protocolset forth in the patent taken by Molecular Probes (16). The details ofour approach are provided in the Supporting Information. For thepresent study, we only used this independently synthesized material toestablish the structure of SOSG.
5,10,15,20-Tetrakis(N-methyl-4-pyridyl)-21H, 23H-porphine(TMPyP, 98%) was purchased from Porphyrin Systems and used asthe independently added photosensitizer in the cell studies. Thephotosensitizer Phenalen-1-one-2-sulfonic acid (PNS), used as astandard for singlet oxygen quantum yield measurements, wassynthesized following a protocol described by Nonell et al. (24).Fluorescein (95%; Sigma Aldrich), bovine serum albumin (BSA, 99%;Sigma Aldrich) and NaN3 (>99%; Merck) were used as received.
ET
OHO
Cl
COOH
HOOC
O OHO O
Cl
COOH
HOOCO
O
weakly ylhgihtnecseroulf fluorescent
O2(a1 g)
ET
SOSG SOSG-EPFigure 1. Formation of the endoperoxide of SOSG (denoted SOSG-EP) upon reaction of SOSG with singlet oxygen. Prior to the reaction withsinglet oxygen, intramolecular electron transfer (ET) quenches the fluorescence from the light-emitting chromophore. Upon reaction with singletoxygen and the formation of the endoperoxide, electron transfer is precluded, and fluorescence is observed. As indicated in Materials and Methods,the SOSG commercially obtained is likely a mixture of the 2,4 and 2,5 dicarboxylic acids, not just the 2,4 isomer shown here.
2 Anita Gollmer et al.
moiety coupled to a light-emitting chromophore. Prior to thereaction with singlet oxygen, emission from the chromophoreis quenched by electron transfer from the adjacent trappingmoiety. Upon reaction with singlet oxygen, however, theresultant oxygen adduct is no longer an efficient intramolec-ular electron donor, and light emission readily occurs.
Two clear advantages of this indirect method to detectsinglet oxygen are (1) the luminescence quantum efficiency ofthe optical probe is comparatively large and (2) emissionoccurs in the visible region of the spectrum where opticaldetectors are very efficient. In turn, excellent optical signalscan be obtained from correspondingly smaller amounts ofsinglet oxygen in a given system. Of course, a limitation ofthis approach is that the probe may also respond to theeffects of reactive oxygen species other than singlet oxygen.Thus, a key aspect of characterizing such systems is toascertain the selectivity of the given probe to singlet oxygen(13–15).
The basic principle and structure of the two-componenttrap-fluorophore system (13) have since been exploited for thedevelopment of a commercially available product: SingletOxygen Sensor Green! (SOSG), Fig. 1 (15). Although thecompany that produces and sells SOSG has not explicitlyassociated a chemical structure with the name SOSG, themolecule shown in Fig. 1 is shown in the company’s patent(16) and, as outlined below, we have independently ascertainedthat this structure is indeed correct.
Although SOSG is a relatively new product, it has alreadybeen used in a number of published studies (17–21). Ofparticular importance, certainly in light of the work presentlydiscussed herein, is the report that SOSG can itself sensitize theproduction of singlet oxygen (22). Although the SOSG-sensi-tized singlet oxygen yields reported in this latter study are small(e.g. FD = 0.006 ± 0.002 upon 355 nm irradiation in aD2O ⁄CH3OH mixture), they are nevertheless not negligible.
In the present study, we further elaborate aspects of SOSGphotophysics and photochemistry, including features of singletoxygen production sensitized by (1) SOSG and (2) theimmediate product of the reaction between SOSG and singletoxygen, the SOSG endoperoxide (denoted SOSG-EP).
Although it has been reported that the behavior and responseof SOSG depends on the wavelength of irradiation (i.e. 355 vs532 nm) (22), we have generally limited our present study to anirradiation wavelength of 420 nm. This specific wavelengthwas chosen, in part, because porphyrins, and many porphyrinderivatives, are often used as singlet oxygen sensitizers, andthese molecules invariably have a pronounced absorption bandin this spectral domain (i.e. the Soret band). Moreover, manyendogenous singlet oxygen sensitizers have an appreciableabsorption cross section at 420 nm (23).
We also present data to indicate that SOSG can beincorporated into a living mammalian cell. This is an obser-vation that, heretofore, has not been reported; indeed, SOSG ismarketed as a cell-impermeable compound (15).
MATERIALS AND METHODSSample preparation. The SOSG used for all experiments reportedherein was purchased from Molecular Probes ⁄ Invitrogen. The purityand exact structure of this SOSG is unknown. In particular, in contrastto what is shown in Fig. 1 (i.e. 2,4-dicarboxylic acid on the phenylgroup), which is based on a structure shown in the Molecular Probespatent (16), it is likely that the material received consists of a mixtureof dicarboxylic acid isomers (see Supporting Information). Thematerial obtained was stored in a desiccator at £)20"C and protectedfrom light. The SOSG solutions used for the experiments wereprepared by first dissolving the compound in methanol (>99%; SigmaAldrich) and then diluting with D2O (>99%; Eurisotop) to yield a97:3 volume-to-volume D2O:CH3OH ratio. The residual alcohol isneeded to ensure complete dissolution of SOSG. Solutions of SOSGwere prepared under dim light (aerated solutions of SOSG arenevertheless reasonably stable upon exposure to ambient light).
Independently, SOSG was synthesized using the general protocolset forth in the patent taken by Molecular Probes (16). The details ofour approach are provided in the Supporting Information. For thepresent study, we only used this independently synthesized material toestablish the structure of SOSG.
5,10,15,20-Tetrakis(N-methyl-4-pyridyl)-21H, 23H-porphine(TMPyP, 98%) was purchased from Porphyrin Systems and used asthe independently added photosensitizer in the cell studies. Thephotosensitizer Phenalen-1-one-2-sulfonic acid (PNS), used as astandard for singlet oxygen quantum yield measurements, wassynthesized following a protocol described by Nonell et al. (24).Fluorescein (95%; Sigma Aldrich), bovine serum albumin (BSA, 99%;Sigma Aldrich) and NaN3 (>99%; Merck) were used as received.
ET
OHO
Cl
COOH
HOOC
O OHO O
Cl
COOH
HOOCO
O
weakly ylhgihtnecseroulf fluorescent
O2(a1 g)
ET
SOSG SOSG-EPFigure 1. Formation of the endoperoxide of SOSG (denoted SOSG-EP) upon reaction of SOSG with singlet oxygen. Prior to the reaction withsinglet oxygen, intramolecular electron transfer (ET) quenches the fluorescence from the light-emitting chromophore. Upon reaction with singletoxygen and the formation of the endoperoxide, electron transfer is precluded, and fluorescence is observed. As indicated in Materials and Methods,the SOSG commercially obtained is likely a mixture of the 2,4 and 2,5 dicarboxylic acids, not just the 2,4 isomer shown here.
2 Anita Gollmer et al.
moiety coupled to a light-emitting chromophore. Prior to thereaction with singlet oxygen, emission from the chromophoreis quenched by electron transfer from the adjacent trappingmoiety. Upon reaction with singlet oxygen, however, theresultant oxygen adduct is no longer an efficient intramolec-ular electron donor, and light emission readily occurs.
Two clear advantages of this indirect method to detectsinglet oxygen are (1) the luminescence quantum efficiency ofthe optical probe is comparatively large and (2) emissionoccurs in the visible region of the spectrum where opticaldetectors are very efficient. In turn, excellent optical signalscan be obtained from correspondingly smaller amounts ofsinglet oxygen in a given system. Of course, a limitation ofthis approach is that the probe may also respond to theeffects of reactive oxygen species other than singlet oxygen.Thus, a key aspect of characterizing such systems is toascertain the selectivity of the given probe to singlet oxygen(13–15).
The basic principle and structure of the two-componenttrap-fluorophore system (13) have since been exploited for thedevelopment of a commercially available product: SingletOxygen Sensor Green! (SOSG), Fig. 1 (15). Although thecompany that produces and sells SOSG has not explicitlyassociated a chemical structure with the name SOSG, themolecule shown in Fig. 1 is shown in the company’s patent(16) and, as outlined below, we have independently ascertainedthat this structure is indeed correct.
Although SOSG is a relatively new product, it has alreadybeen used in a number of published studies (17–21). Ofparticular importance, certainly in light of the work presentlydiscussed herein, is the report that SOSG can itself sensitize theproduction of singlet oxygen (22). Although the SOSG-sensi-tized singlet oxygen yields reported in this latter study are small(e.g. FD = 0.006 ± 0.002 upon 355 nm irradiation in aD2O ⁄CH3OH mixture), they are nevertheless not negligible.
In the present study, we further elaborate aspects of SOSGphotophysics and photochemistry, including features of singletoxygen production sensitized by (1) SOSG and (2) theimmediate product of the reaction between SOSG and singletoxygen, the SOSG endoperoxide (denoted SOSG-EP).
Although it has been reported that the behavior and responseof SOSG depends on the wavelength of irradiation (i.e. 355 vs532 nm) (22), we have generally limited our present study to anirradiation wavelength of 420 nm. This specific wavelengthwas chosen, in part, because porphyrins, and many porphyrinderivatives, are often used as singlet oxygen sensitizers, andthese molecules invariably have a pronounced absorption bandin this spectral domain (i.e. the Soret band). Moreover, manyendogenous singlet oxygen sensitizers have an appreciableabsorption cross section at 420 nm (23).
We also present data to indicate that SOSG can beincorporated into a living mammalian cell. This is an obser-vation that, heretofore, has not been reported; indeed, SOSG ismarketed as a cell-impermeable compound (15).
MATERIALS AND METHODSSample preparation. The SOSG used for all experiments reportedherein was purchased from Molecular Probes ⁄ Invitrogen. The purityand exact structure of this SOSG is unknown. In particular, in contrastto what is shown in Fig. 1 (i.e. 2,4-dicarboxylic acid on the phenylgroup), which is based on a structure shown in the Molecular Probespatent (16), it is likely that the material received consists of a mixtureof dicarboxylic acid isomers (see Supporting Information). Thematerial obtained was stored in a desiccator at £)20"C and protectedfrom light. The SOSG solutions used for the experiments wereprepared by first dissolving the compound in methanol (>99%; SigmaAldrich) and then diluting with D2O (>99%; Eurisotop) to yield a97:3 volume-to-volume D2O:CH3OH ratio. The residual alcohol isneeded to ensure complete dissolution of SOSG. Solutions of SOSGwere prepared under dim light (aerated solutions of SOSG arenevertheless reasonably stable upon exposure to ambient light).
Independently, SOSG was synthesized using the general protocolset forth in the patent taken by Molecular Probes (16). The details ofour approach are provided in the Supporting Information. For thepresent study, we only used this independently synthesized material toestablish the structure of SOSG.
5,10,15,20-Tetrakis(N-methyl-4-pyridyl)-21H, 23H-porphine(TMPyP, 98%) was purchased from Porphyrin Systems and used asthe independently added photosensitizer in the cell studies. Thephotosensitizer Phenalen-1-one-2-sulfonic acid (PNS), used as astandard for singlet oxygen quantum yield measurements, wassynthesized following a protocol described by Nonell et al. (24).Fluorescein (95%; Sigma Aldrich), bovine serum albumin (BSA, 99%;Sigma Aldrich) and NaN3 (>99%; Merck) were used as received.
ET
OHO
Cl
COOH
HOOC
O OHO O
Cl
COOH
HOOCO
O
weakly ylhgihtnecseroulf fluorescent
O2(a1 g)
ET
SOSG SOSG-EPFigure 1. Formation of the endoperoxide of SOSG (denoted SOSG-EP) upon reaction of SOSG with singlet oxygen. Prior to the reaction withsinglet oxygen, intramolecular electron transfer (ET) quenches the fluorescence from the light-emitting chromophore. Upon reaction with singletoxygen and the formation of the endoperoxide, electron transfer is precluded, and fluorescence is observed. As indicated in Materials and Methods,the SOSG commercially obtained is likely a mixture of the 2,4 and 2,5 dicarboxylic acids, not just the 2,4 isomer shown here.
2 Anita Gollmer et al.
moiety coupled to a light-emitting chromophore. Prior to thereaction with singlet oxygen, emission from the chromophoreis quenched by electron transfer from the adjacent trappingmoiety. Upon reaction with singlet oxygen, however, theresultant oxygen adduct is no longer an efficient intramolec-ular electron donor, and light emission readily occurs.
Two clear advantages of this indirect method to detectsinglet oxygen are (1) the luminescence quantum efficiency ofthe optical probe is comparatively large and (2) emissionoccurs in the visible region of the spectrum where opticaldetectors are very efficient. In turn, excellent optical signalscan be obtained from correspondingly smaller amounts ofsinglet oxygen in a given system. Of course, a limitation ofthis approach is that the probe may also respond to theeffects of reactive oxygen species other than singlet oxygen.Thus, a key aspect of characterizing such systems is toascertain the selectivity of the given probe to singlet oxygen(13–15).
The basic principle and structure of the two-componenttrap-fluorophore system (13) have since been exploited for thedevelopment of a commercially available product: SingletOxygen Sensor Green! (SOSG), Fig. 1 (15). Although thecompany that produces and sells SOSG has not explicitlyassociated a chemical structure with the name SOSG, themolecule shown in Fig. 1 is shown in the company’s patent(16) and, as outlined below, we have independently ascertainedthat this structure is indeed correct.
Although SOSG is a relatively new product, it has alreadybeen used in a number of published studies (17–21). Ofparticular importance, certainly in light of the work presentlydiscussed herein, is the report that SOSG can itself sensitize theproduction of singlet oxygen (22). Although the SOSG-sensi-tized singlet oxygen yields reported in this latter study are small(e.g. FD = 0.006 ± 0.002 upon 355 nm irradiation in aD2O ⁄CH3OH mixture), they are nevertheless not negligible.
In the present study, we further elaborate aspects of SOSGphotophysics and photochemistry, including features of singletoxygen production sensitized by (1) SOSG and (2) theimmediate product of the reaction between SOSG and singletoxygen, the SOSG endoperoxide (denoted SOSG-EP).
Although it has been reported that the behavior and responseof SOSG depends on the wavelength of irradiation (i.e. 355 vs532 nm) (22), we have generally limited our present study to anirradiation wavelength of 420 nm. This specific wavelengthwas chosen, in part, because porphyrins, and many porphyrinderivatives, are often used as singlet oxygen sensitizers, andthese molecules invariably have a pronounced absorption bandin this spectral domain (i.e. the Soret band). Moreover, manyendogenous singlet oxygen sensitizers have an appreciableabsorption cross section at 420 nm (23).
We also present data to indicate that SOSG can beincorporated into a living mammalian cell. This is an obser-vation that, heretofore, has not been reported; indeed, SOSG ismarketed as a cell-impermeable compound (15).
MATERIALS AND METHODSSample preparation. The SOSG used for all experiments reportedherein was purchased from Molecular Probes ⁄ Invitrogen. The purityand exact structure of this SOSG is unknown. In particular, in contrastto what is shown in Fig. 1 (i.e. 2,4-dicarboxylic acid on the phenylgroup), which is based on a structure shown in the Molecular Probespatent (16), it is likely that the material received consists of a mixtureof dicarboxylic acid isomers (see Supporting Information). Thematerial obtained was stored in a desiccator at £)20"C and protectedfrom light. The SOSG solutions used for the experiments wereprepared by first dissolving the compound in methanol (>99%; SigmaAldrich) and then diluting with D2O (>99%; Eurisotop) to yield a97:3 volume-to-volume D2O:CH3OH ratio. The residual alcohol isneeded to ensure complete dissolution of SOSG. Solutions of SOSGwere prepared under dim light (aerated solutions of SOSG arenevertheless reasonably stable upon exposure to ambient light).
Independently, SOSG was synthesized using the general protocolset forth in the patent taken by Molecular Probes (16). The details ofour approach are provided in the Supporting Information. For thepresent study, we only used this independently synthesized material toestablish the structure of SOSG.
5,10,15,20-Tetrakis(N-methyl-4-pyridyl)-21H, 23H-porphine(TMPyP, 98%) was purchased from Porphyrin Systems and used asthe independently added photosensitizer in the cell studies. Thephotosensitizer Phenalen-1-one-2-sulfonic acid (PNS), used as astandard for singlet oxygen quantum yield measurements, wassynthesized following a protocol described by Nonell et al. (24).Fluorescein (95%; Sigma Aldrich), bovine serum albumin (BSA, 99%;Sigma Aldrich) and NaN3 (>99%; Merck) were used as received.
ET
OHO
Cl
COOH
HOOC
O OHO O
Cl
COOH
HOOCO
O
weakly ylhgihtnecseroulf fluorescent
O2(a1 g)
ET
SOSG SOSG-EPFigure 1. Formation of the endoperoxide of SOSG (denoted SOSG-EP) upon reaction of SOSG with singlet oxygen. Prior to the reaction withsinglet oxygen, intramolecular electron transfer (ET) quenches the fluorescence from the light-emitting chromophore. Upon reaction with singletoxygen and the formation of the endoperoxide, electron transfer is precluded, and fluorescence is observed. As indicated in Materials and Methods,the SOSG commercially obtained is likely a mixture of the 2,4 and 2,5 dicarboxylic acids, not just the 2,4 isomer shown here.
2 Anita Gollmer et al.
Figure 4: Oxidation of SOSG by 1O2. Adapted from [83].
2.1.1.2 ROS Standard Generators
Quantification of ROS production detected via probe compounds is possible
using known ROS generators. Often, these are dyes or other chromaphores. For
example, rose bengal exposed to UV light at 365nm will generate singlet oxygen with a
quantum yield of 0.75 [95-97, 103]. By measuring UV fluence and the absorbance of the
sensitizer at a given concentration, the 1O2 generation rate can be calculated. This can
then be combined with the response of the probe compound to create a standard
response curve. Similarly, xanthine combined with xanthine oxidase is a known
superoxide generator [98], while hydrogen peroxide can be used to generate hydroxyl
radicals [99, 100].
16
2.1.1.3 ROS Quenchers
The use of quenching agents can help to isolate the effect of a single species when
investigating the ROS mediated inactivation of bacteria or viruses. Because the
pathways though which ROS can react are often complex and interrelated, the use of
compounds that specifically react with a given radical at a rate large enough to
outcompete other reactions can greatly clarify observations [70, 71, 104]. Frequently
employed quenching agents are typically biologically synthesized molecules that are
produced to protect cells from the oxidizing and inflammatory effect of free radicals
such as superoxide dismutase for superoxide, or β-carotene and histidine for singlet
oxygen [90-92, 105]. Care should be taken in selecting an appropriate quencher when
studying biological systems as some options (e.g. N-acetyl-cysteine) are biologically safe
while others (e.g. sodium azide) have known toxicity, though both are effective �OH
quenchers [93, 94, 106].
2.2 Photoreactive Nanoparticles
Photoreactive nanoparticles may be semiconductors, often metal oxides (e.g.
WO3, SrTiO3, Fe2O3, ZnO) or metal sulfides (CdS, ZnS), or carbon based NPs (e.g.
fullerene nanoparticles (FNPs), carbon nanotubes (CNTs), and graphene) [16, 107, 108].
In semiconductors, the difference in energy levels between the highest occupied
molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) is
sufficient that electrons are unable to freely occupy higher orbitals due to relatively
17
insignificant changes in energy (as is the case for conductive metals), but not so great
that the band gap between the HOMO and LUMO is insurmountable (as is the case for
insulators).
As the difference in the HOMO and LUMO determines the energy necessary to
promote an electron into the conduction band, a small band gap is desirable for
photocatalysis, which enables excitation by photons of greater wavelength. The
absorption of a photon with energy greater than or equal to the band gap will promote
an electron from the HOMO, across the band gap into the LUMO, leaving behind a
photogenerated hole, h+. From here, the presence of vacant orbitals permits the
migration of the photoëxcited electron through the conduction band, while the
photogenerated hole is free to migrate through the valence band [40]. Once the electron-
hole pair has been generated, they may recombine, releasing energy via fluorescence.
Alternatively, migration to the surface of the sensitizer and interaction of h+ with water
or the electron with oxygen can produce hydroxyl radicals and superoxide, respectively
(Figure 7).
To be practically useful, the semiconductor should be insoluble and inert in
water. Additionally, the band gap should span the redox potential of water in order to
permit �OH production [108]. While TiO2 is often selected for its inert nature and
relative abundance, the band gap of 3.2 and 3.0 V for anatase and rutile, respectively
means that light with wavelengths of less than 415 nm for rutile, and 385 nm for anatase,
18
is necessary for photoëxcitation [16]. This is a significant limiting factor, with
implications for cost and practicality. When considering TiO2 for photocatalytic
processes, it necessitates the use of UV lamps, and considerable research has
investigated semiconductor doping in an attempt to decrease the band gap so that less
expensive visible light (i.e. solar) sources can be effectively employed [41, 109, 110].
While this is an extensive and active field of research, it falls outside the scope of work
presented in this dissertation.
Similarly, for carbonaceous NPs the absorption of light excites an electron into
the LUMO [111, 112]. Relaxation can take the form of radiative decay (phosphorescence),
increased vibration (heat exchange), or chemical transformation. Chemical
transformations will either take place through degradation, typically oxidation in the
presence of dissolved oxygen, or catalysis in which energy is transferred through either
type I or type II sensitization (Figure 6) [40].
In determining the efficacy of a sensitizer, the quantum yield, the ratio of
absorbed photons resulting in this latter pathway to all non-catalytic pathways, is
employed. This propensity for chemical reactions ranges from zero (no reaction) to unity
(all absorbed photons result in a reaction) [49]. Knowing this, the incident light intensity,
and the absorptivity of the compound, one can calculate the ROS production from a
given concentration of a photocatalyst.
19
2.2.1 Fullerene Nanoparticles
Fullerene, first described in the literature in 1985 [113], is a carbon allotrope most
commonly composed of 60 carbon atoms (C60) forming interlocked hexagonal and
pentagonal rings in sp2 hybridization, Figure 5 (a). The extensive π-electron bonding
throughout the carbon cage gives rise to a number of interesting and potentially useful
parameters, including strength, resistance to degradation, photocatalytic ability, and, in
a complicating factor when considering their employment in aqueous systems,
significant hydrophobicity [26, 114, 115].
Due to this hydrophobic nature of C60, larger aggregates composed of “n”
number of individual monomers, referred to as nC60, are generally formed in aqueous
suspensions. In an attempt to minimize the size and number of monomers within
aggregates, multiple methods of dispersion have been attempted including long term
stirring, solvent exchange, and high power ultrasonication, to varying degrees of success
and particle alteration [116, 117]. Though solvent exchange, often with tetrahydrofuran,
was initially utilized for FNP dispersions, toxicity measurements were complicated by
residual solvent [118-120]. Additionally, long term stirring and sonication have been
observed to hydroxylate FNP aggregates in suspension [121, 122].
Additionally, derivatization of the carbon cage, wherein some of the π-electron
bonding is disrupted to allow for the addition of functional groups to the outside of the
cage, has been employed to increase solubility. Frequently this is instigated through the
20
addition of hydroxyl, amine, or other small functional groups, though large side chains
have been used both for aqueous stabilization, cross-linking with polymer substrates, or
vehicles for drug delivery [123-126]. The most commonly studied functionalized
fullerene, nominally with 24 hydroxyl groups, is fullerol (C60(OH)24) Figure 5 (b) [22, 78,
127-129]. Functionalization not only alters solubility but impacts many other physical
and photochemical properties of the FNP. OH
OH
HO
HO
NH2
H2N
NH2
NH2
H2N
OHHO
HO HO
OH
OH
OH
HOHO
OH
OH
OH
OH
HO
HO
NH2
H2N
NH2
NH2
H2N
OHHO
HO HO
OH
OH
OH
HOHO
OH
OH
a) b)
Figure 5: (a) C60, fullerene and (b) C60(OH)24, fullerol.
2.2.1.1 FNP Reactivity
Fullerenes are efficient photosensitizers capable in solution of transferring
energy from absorbed light to dissolved oxygen [112, 130, 131]. This occurs as the
absorption of UV light propagates an electron from the ground state to the excited
singlet state. This short lived energy state rapidly (order of nanoseconds) decays to the
relatively stable excited triplet state (tens of microseconds) where energy can further be
transferred to dissolved oxygen in solution, creating 1O2 as shown in Figure 6 [111].
21
Alternatively, if an electron donor is present, the excited triplet state may abstract that
electron, producing a radical anion, after which the reaction with dissolved oxygen will
produce superoxide (O2�-) [130-132].
Fullerene+Sensi;za;on+
Figure+4.4+Photosensi;za;on+of+C60.+Adapted+from+Brunet+et+al.78++
Figure 6: Photosensitization of and ROS production by C60. Adapted from [81].
Nucleophilic addition reactions are capable of producing a vast number of
products by adding functional groups to the fullerene cage [123]. The functionalization
of C60 has been shown to change singlet state absorption and decay and decrease
photochemical ability [80, 133]. However, most investigations have involved the
addition of large functional groups compared to the relatively small carbon cage [125,
134-136]. This trend has been continued more recently through investigations of singlet
production and virus inactivation after large chain additions [32, 137, 138]. The presence
of simple, low molecular weight additions has not yet been thoroughly investigated.
Given the photoreactivity of FNPs, there has been interest in their antimicrobial
and anti-viral capabilities, both as a targeted application of FNPs or an unintended
22
consequence of environmental release [52]. Badireddy et al. investigated the inactivation
of MS2 virus by fullerol and observed that viral inactivation was due predominantly to
singlet oxygen production [128]. Generally, no inactivation is seen under dark
conditions for sonicated aqueous suspensions [139].
2.2.1.2 FNP Physical Transformations
Owing to the hydrophobicity of C60, the size distributions of aqueous dispersions
are often highly heterogeneous. Chae et al. demonstrated the relationship between
aggregate diameter and characteristics such as pH, ORP, and the coefficient of
attachment efficiency (α) by filtering a suspension nC60 to different pore sizes [121]. The
conclusion was that the stability of a fullerene or aggregate changes as a result of surface
hydroxylation during suspension preparation.
This work built upon previous studies looking at the possibility of surface
transformations either during or after suspension preparation. Brant et al., in looking at
the effect of varying solvent exchange dispersion methods, found through light
scattering that the internal structure, as well as aggregate size, differed based on
dispersion methods [116]. These findings were attributed to varying degrees of
aggregate stabilization due to electrostatic and hydration interactions, a conclusion that
was supported by later studies looking at the irreversible hydration and hydroxylation
of nC60 aggregates during sonication as evidenced by FTIR and NMR [122]. The reaction
of nC60 with hydroxyl radicals produced via cavitation was proposed to be the primary
23
mechanism for this hydroxylation, which was further supported by the work of Lee et
al. using gamma radiation to expose C60 suspensions to hydroxyl radicals [140]. Surface
hydroxylation was also observed through the evolution of UV-vis absorption spectra by
Chang et al. during long-term stirring of C60 in aqueous suspensions without any solvent
exchange [117]. Together, these studies point to the changes that necessarily occur to
pristine, powdered C60 in creating quasi-stable aqueous dispersions.
Furthermore, UV irradiation of prepared C60 suspensions was shown by Lee et al.
to alter the fullerenes into mono- and di-oxygenated products [24]. Others have shown
similar UV-mediated transformations of fullerene [23, 129, 141]. Hwang et al. detailed
changes in suspension characteristics over 21 days, with transformations likely
continuing afterwards [142].
2.2.2 Titanium Dioxide
TiO2 is a semiconducting metal oxide with three different crystalline phases:
anatase, brookite, and rutile. Of these, rutile is the most stable, while brookite is the least
and readily transforms to rutile [143]. Due to it’s photoelectronic characteristics, TiO2 has
widely been studied for many applications, including pigmentation, advanced oxidation
processes, photovoltaics, and biomedical therapies [15, 144-147].
2.2.2.1 TiO2 Reactivity
Of the three crystalline phases of TiO2, anatase shows the greatest photoactivity
[146, 148]. Rutile appears to be a generally less efficient photocatalyst than anatase for
24
the majority of compounds, but has shown high photocatalytic ability for some [15, 149].
Brookite is the least often employed, owing partly to its difficulty to prepare and purify
[150, 151]. Additionally, Bakardjieva et al. found the inclusion of brookite did not
markedly improve reactivity when testing various combinations of crystalline phases
[152, 153].
As a semiconductor, the absorption of light greater than or equal to the band gap
will promote an electron from the HOMO to the LUMO, leaving behind a positive hole.
For undoped TiO2 in the anatase crystalline phase, this occurs at approximately 385nm
(corresponding to a band gap of 3.2 V) [16]. As shown in Figure 7, once separated, both
the electron and the hole may migrate to the surface of the particle where they are
capable of undergoing a number of reactions, either reducing or oxidizing compounds
present on the crystalline surface [40, 154, 155]. Additionally, they may recombine,
releasing the absorbed energy through photoluminescence or non-radiative decay in
processes that produce no chemical transformation. For ROS forming pathways, the
electron, after migrating to the surface, is capable of reducing dissolved oxygen to
produce superoxide, while the hole is able to abstract an electron from hydroxide,
producing the highly oxidizing hydroxyl radical [15, 156].
While much of the environmental TiO2 research focuses on compound
degradation [40, 107, 146, 157-160], photoactivated TiO2 has proven effective at bacteria
and virus inactivation, with �OH playing a predominant role in toxicity [71, 161-163].
25
With TiO2 increasingly used in consumer products manufacturing (e.g. sunscreens,
polymers and paints, cements) [164-168], the release of TiO2 into the environment will
continue to increase [169-171]. As such, applications should also take into account the
potential for unintended biological impacts resulting from these products [172-174].
TiO2+Photocatalysis+
−OH$
OH$
O2$
O2−$
Figure+4.5+TiO2+ROS+produc;on+schema;c.+Adapted+from+Linsebigler+et+al.28++
Figure 7: Photosensitization of and ROS production by TiO2. Adapted from [40].
2.2.2.2 P25 TiO2
P25, manufactured by Evonik (Germany), is the most commonly cited
commercial TiO2 product in the literature and consists of a combination of the rutile and
anatase phases fused together in roughly a 20:80 ratio [175]. The individual particle size
is nominally 25nm with a BET surface area of about 55 m2 g-1 [15, 175, 176], however the
26
fused nature of the particles results in ultrasonicated dispersions of hundreds of
nanometers [177, 178]. P25 has been observed to produce greater �OH than similar
dispersions of anatase alone (Figure 8), attributed in part to the formation of trapping
sites at the rutile-anatase junctions [155]. Bakardjieva et al. also demonstrated that the
ratio of anatase to rutile impacts particle size, pore size and, ultimately photoreactivity
[153]. None of their samples bettered P25 at the degradation of 4-chlorophenol, though
similar ratios behaved with much the same effectiveness.
320 340 360 380 400 420 440 460 4800
100
200
300
400
500
600
700
800
900
Emission Wavelength (nm)
Fluo
resc
ence
(FSU
)
P25ArcosAldrichNanoAmorP25 Dark
Figure 8: �OH production of various TiO2 suspensions at 20ppm, measured by TA and graphed as fluorescence intensity vs. emission wavelength after 5 minutes of
UV-A exposure. Unpublished data.
27
2.3 Particle Stability and Aggregation
2.3.1 Particle Attachment
When two particles collide in a fluid, the likelihood that they will attach and
produce an aggregate is indicated by the parameter α, the coefficient of attachment
efficiency. This “stickiness coefficient” is the ratio of particle attachment occurrences to
particle collisions, ranging from zero (no attachment) to unity (every collision results in
attachment), and is governed primarily by the surface chemistry of the particle [179].
Van der Waal’s forces, the weakly attractive forces between two particles, are easily
overcome by electrostatic repulsion in the bulk. As two similarly charged particles come
towards each other, the net repulsive force increases. However, because the charge on
the surface of a particle is finite, and van der Waal’s attraction forces continually
increase as the separation distance drops to zero, the net force, considering these two
phenomena, will typically be attractive at a sufficiently small separation distance. Thus
particles brought closely enough together will continue to approach each other until the
electron clouds of individual atoms begin to overlap, as dictated by Born repulsion.
DLVO theory quantitatively describes this net attractive or repulsive force for particles
of varying separation distances (Figure 9) [180, 181].
28
DLVO+Theory+Schema;c+
Separation Distance
Pote
ntia
l Ene
rgy
Electosta;c+repulsion+
Van+der+Waal’s+Afrac;on+
DLVO+
Secondary+Minimum+
Barrier+to+Aggrega;on+
Figure 9: DLVO schematic illustrating the sum (solid black line) of electrostatic repulsion (blue dashed line) and van der Waal’s attraction (red dashed
line) as a function of separation distance.
The magnitude of electrostatic repulsive forces depends on the surface charge,
conferred by lattice structure, surface defects in the crystalline phase, and adsorbed
species or functional groups. Effectively, this charge on the surface of a particle draws
near a fixed layer of ions of opposite charge existing in solution as a counterbalance, the
Stern layer. The concentration of counter-ions decreases with distance from the surface
until it reaches an effective balance between positive and negative at which point
particles no longer feel the impact of each other. The distance over which this occurs, the
Debeye length (κ-1), depends on the magnitude of the charge on the surface, however, as
29
is seen in Equation 5, for a given surface charge, this length can be decreased though
charge screening, accomplished by increasing the ionic strength of the solution:
κ −1 =εkBT2NAe
−2I (Eq 5.)
Here ε is the permittivity of the system, kB is the Boltzman constant, T is temperature,
NA is Avagadro’s number, e is the charge of an electron, and I is the ionic strength of the
solution. When measuring particle stability via electrophoretic mobility, the value
calculated as the zeta (ζ) potential is the net charge at the slipping plane rather than the
charge on the surface itself or at the Stern layer. Together, these regimes make up the
electrical double layer of a particle [182].
Additionally, adsorbed species or functional groups on the surface of a particle
can increase stability through steric hindrance, preventing other particles from
approaching close enough to overcome the repulsive force [29, 183]. This often occurs to
particles naturally from environmental organic matter or biological molecules [184-187]
and can be engineered through the use of various surfactants and stabilizers [25, 188,
189]. Under these cases, classical DLVO theory breaks down, and a number of equations
have been derived to describe extended DLVO behavior based on the particle
characteristics [29, 189-191].
30
2.3.2 Rectilinear Transport and Particle Collisions
For two particles to aggregate, they must first be brought together. The process
by which particles come into contact is governed by both the characteristics of the
particles and the solution. In a rectilinear transport model, interactions between the
stream lines of approaching particles are ignored and transport is assumed to follow
straight flow lines. Three collision mechanisms are typically considered: fluid shear
(kSH), Brownian motion (kB), and differential settling (kDS) such that
β = kSH + kB + kDS (Eq 6.)
where β is the combined rate of collisions between particles in units of per number
concentration per time.
Fluid shear is governed by the velocity gradients within the fluid and produces a
ballistic transport mechanism. Particles carried along a path of higher velocity will
overtake those travelling at a lower velocity denoted in Equation 7 where di and dj are
the diameters of particles i and j, respectively, and G is the mean velocity gradient
produced by mixing in units of second-1.
kSH (i, j) =(di − dj )
3G6
(Eq 7.)
Brownian motion is determined by the viscosity of the fluid, µ, and the internal
energy (i.e. temperature) of the particle. The contribution from Brownian motion
dominates for small particles under quiescent conditions.
31
kB (i, j) =2kBT3µ
1di+1dj
!
"##
$
%&& di + dj( ) (Eq 8.)
Differential settling (Equation 9) applies to particles or aggregates of varying
size, in which the settling rate of larger or denser particles causes them to travel more
quickly, overcoming smaller or less dense particles in their path. Here, g is the force of
gravity.
kDS (i, j) =πg(ρi − ρ j )72µ
(di + dj )2 (di
2 − dj2 ) (Eq 9.)
The overall rate of aggregation is a combination of both particle attachment (α)
and transport (β), described by the Smoluchowski equation:
dNk
dt=12α βijNiN ji+ j=k∑ −αNk βikNii=1
∞
∑ (Eq 10.)
Here dNk
dt describes the accumulation of particles of a given size, k, with generation due
to the combination of smaller size classes, i and j, and loss due to aggregation of size
class k to larger sizes [192, 193].
The rate of aggregation impacts not only the size of aggregates, but also their
structure. The size of an aggregate and number of primary particles can be related to the
density by the fractal dimension, Df, which constitutes a description of aggregate self-
similarity across length scales [194].
32
2.3.3 Fractal Dimension
If particles are fully destabilized (α=1), aggregation is considered diffusion
limited (DLA) (Figure 10 (a)). Under DLA, the barrier to attachment is effectively
nonexistent, and a particle will attach to the first surface with which it collides, rapidly
producing relatively large, open aggregates with fractal dimensions around 1.8 [195].
Conversely, smaller, denser aggregates are produced if aggregation is slow and multiple
collisions are necessary before attachment occurs. This latter condition is termed reaction
limited aggregation (RLA) and is characterized by Df values of 2.1 or greater (Figure 10
(b)) [195, 196]. Particles brought together by fluid flow or deposition (in contrast to
Brownian motion) undergo ballistic aggregation, resulting in aggregates for which the
Df generally lies in between values for DLA and RLA aggregation [196, 197].
180 J. Zhang, J. Buffle/Colloids Surfaces A: Physicochem. Eng. Aspects 107 (1996) 175-187
for reaction limited aggregation [22,23]. The
experimental curve at 55 mM, shown in Fig. la,
was also exponential-like when viewed over
extended aggregation time. However, at even lower
KC1 concentrations, due to slow aggregation kinet-
ics, the shape of the curves was difficult to deter-
mine within 24 h. It is very interesting to note that
the transition from a concave to a convex shaped
curves occurred in a very narrow KC1 concen-
tration range of 65-75 mM. The initial rapid
increase in the aggregate size in experiments at
[ K C 1 ] > 6 5 m M is probably due to orthokinetic
coagulation during mixing.
4.2. Dynamic scaling
The fractal dimension of diffusion limited aggre-
gates may be computed from Eq. (15) if the number
average radius of gyration is known. But the
parameter determined by PCS is the z-average
1000
E r "
100
Slope=0.534
I 10
100
Time/rain
Fig. 2. Logarithmic plot of the mean hydrodynamic size versus
aggregation time for DLA ([KC1] =200 mM).
(a)
(b)
Fig. 3. Transmission electromicrographs of typical aggregates
formed in (a) DLA regime for [KCI]=0.3 M, and (b) RLA
regime for [KC1] = 55 mM.
hydrodynamic size. In fact, these two parameters
are approximately proportional under our experi-
mental conditions [ 12]. Thus the fractal dimension
may be estimated from the logarithmic plots of
z-average size against time. A typical plot is shown
in Fig. 2 which is linear over the size range
200-1200 nm ([KC1] =0.2 M). At larger sizes, the
slope decreases, probably due to the sedimentation
of aggregates. Eight experiments in the DLA
180
J. Zh
ang,
J. B
uffle
/Col
loid
s Su
rface
s A: P
hysic
oche
m. E
ng. A
spec
ts 10
7 (1
996)
175
-187
for
reac
tion
limite
d ag
greg
atio
n [2
2,23
]. Th
e
expe
rimen
tal
curv
e at
55
mM
, sh
own
in F
ig. l
a,
was
al
so
expo
nent
ial-l
ike
whe
n vi
ewed
ov
er
exte
nded
agg
rega
tion
time.
How
ever
, at e
ven
low
er
KC1
con
cent
ratio
ns, d
ue to
slo
w a
ggre
gatio
n ki
net-
ics, t
he s
hape
of t
he c
urve
s w
as d
iffic
ult t
o de
ter-
min
e w
ithin
24
h. I
t is v
ery
inte
rest
ing
to n
ote
that
the
trans
ition
fro
m a
con
cave
to
a co
nvex
sha
ped
curv
es o
ccur
red
in a
ver
y na
rrow
KC1
con
cen-
trat
ion
rang
e of
65-
75 m
M.
The
initi
al
rapi
d
incr
ease
in
the
aggr
egat
e siz
e in
exp
erim
ents
at
[KC
1]>6
5mM
is
prob
ably
due
to
orth
okin
etic
coag
ulat
ion
duri
ng m
ixin
g.
4.2.
Dyn
amic
scal
ing
The
frac
tal d
imen
sion
of d
iffus
ion
limite
d ag
gre-
gate
s may
be
com
pute
d fr
om E
q. (1
5) if
the
num
ber
aver
age
radi
us
of g
yrat
ion
is kn
own.
Bu
t th
e
para
met
er d
eter
min
ed b
y PC
S is
the
z-av
erag
e
1000 E r"
100
Slop
e=0.
534
I
10
100
Time/rain
Fig.
2.
Loga
rithm
ic p
lot o
f the
mea
n hy
drod
ynam
ic si
ze v
ersu
s
aggr
egat
ion
time
for D
LA (
[KC1
] =20
0 m
M).
(a)
(b) Fi
g. 3
. Tr
ansm
issi
on e
lect
rom
icro
grap
hs o
f ty
pica
l ag
greg
ates
form
ed i
n (a
) D
LA r
egim
e fo
r [K
CI]
=0.3
M,
and
(b)
RLA
regi
me
for
[KC1
] = 5
5 m
M.
hydr
odyn
amic
size
. In
fac
t, th
ese
two
para
met
ers
are
appr
oxim
atel
y pr
opor
tiona
l un
der
our
expe
ri-
men
tal c
ondi
tions
[ 12
]. Th
us th
e fr
acta
l dim
ensi
on
may
be
estim
ated
fro
m t
he l
ogar
ithm
ic p
lots
of
z-av
erag
e siz
e ag
ains
t tim
e. A
typi
cal p
lot i
s sh
own
in
Fig.
2
whi
ch
is lin
ear
over
th
e siz
e ra
nge
200-
1200
nm
([K
C1] =
0.2
M).
At l
arge
r siz
es, t
he
slop
e de
crea
ses,
prob
ably
due
to th
e se
dim
enta
tion
of a
ggre
gate
s. Ei
ght
expe
rimen
ts
in
the
DLA
(a) (b)
DLA Df = 1.68 RLA Df = 2.20
Hematite aggregates, primary particle size: 25nm+/-2nm
FromZhang1995
Figure 10: TEM images of 25nm hematite NPs aggregated under (a) DLA (Df = 1.68) and (b) RLA (Df = 2.20) schemes. Adapted from [198].
33
Measurements of Df consistent with predictions for both RLA and DLA
aggregation schemes have been observed for TiO2 [177, 199], fullerenes [200], and
several other NPs in unmixed systems [198, 201-203]. In mixed systems, loose aggregates
are more susceptible to shear forces, resulting in breakup and reordering. The process of
re-aggregation produces denser aggregates and increases Df [204, 205].
2.3.4 Changes to the Smoluchowski Equation Due to the Fractal Nature of Aggregates
Among the simplifications made by Smoluchowski is that flocculation occurs
between coalescing spheres (particle volume is constant) without taking into account
fluid flow around the particles (rectilinear model) [179]. In fact, as aggregation between
non-coalescing particles occurs, the aggregate volume fraction increases as fluid is
included in the aggregate. In addition to curvilinear effects observed between two
approaching solid particles, Adler demonstrated the compression of flow lines near the
surface of porous particles as fluid flows around and through the particle [206]. The
ability of streamlines to flow through aggregates was further supported by the
recognition from Li and Ganczarczyk that naturally forming aggregates exhibit a fractal
nature [207]. Chellam and Wiesner demonstrated mathematically that this fractal nature
has a direct impact on floc permeability, indicating that very loose aggregates approach
the behavior of the rectilinear model, while for dense aggregates, the curvilinear model
better describes transport [208]. Ultimately however, the rectilinear and curvilinear
34
models bound the possibilities for aggregate interactions, with true aggregate behavior
falling somewhere in between.
The primary implication when considering fractal aggregates is when two
aggregates combine, the resulting occupied volume is greater than the sum of the two
initial volumes, with the exception of when Df = 3 [179]. Thus, the porosity of aggregates
is no longer fixed, but rather increases with radial distance from the center of the
aggregate. Veerapeneni and Wiesner, continuing the work by Chellam and Wiesner,
demonstrated this by dividing aggregates into multiple shells of decreasing porosity.
The fluid collection efficiency, the amount of fluid that passes through an aggregate
divided by the amount that approaches the aggregate, of an aggregate decreases with
increasing fractal dimension, effectively pushing streamlines to the outside of the
aggregate, as illustrated in Figure 11 [209].
Figure 11: Schematic of rectilinear (far left) and curvilinear (far right) transport models alongside calculations of flow lines in fractal aggregates with radially varying
permeability. Adapted from [210] and [209].
35
Practically, the curvilinear model drastically reduces the relative contribution of
mixing in flocculation [210]. The inability of theoretical settling rates for impenetrable
spheres to mirror experimental data has been demonstrated [211, 212]. Porosity
decreases the drag force slowing the particle’s fall, and experimental velocities of
settling aggregates are often orders of magnitude greater than would be expected from
Stokes’ law for solid spheres. Furthermore, the attachment a of small particles by a large
aggregate is overestimated if only aggregate diameter is considered, as flow lines that
pass through large pores between clusters in an aggregate often carry small monomers
with them unaffected [212]. By including fluid collection efficiency and the effective
permeability of fractal aggregates, equations for shear and differential settling which
better reflect experimental observations, as well as breakup due to shear stresses have
been determined [179].
The fractal nature also determines the susceptibility of aggregates to shear
stresses and helps determine the stable particle size distribution in a mixed system [213].
Breakup for denser aggregates occurs through the loss of monomers from the surface
whereas lose aggregates are more susceptible to more dynamic aggregate fracturing
[179, 214].
2.3.5 Measuring Fractal Dimension via Light Scattering
The fractal dimension of aggregates can be determined by studies employing
light, x-ray, or neutron scattering [215]. In each case, the approach is the similar; the
36
intensity of scattered light depends on the scattering of the primary particles, P(q), and
the structure of the aggregate, S(q) according to Equation 11,
I(q)∝P(q)S(q) (Eq 11.)
Where q is the scattering vector in units of length-1, given by
q = 4πnλsin θ
2!
"#$
%& (Eq 12.)
Different length scales can be probed by changing the angle of detection, θ . Depending
on the size of primary particles and of the aggregates, these length scales will fall under
one of several regimes (Figure 12, Table 2).
Table 2: Light scattering regimes of fractal aggregates.
Scattering in the Rayleigh regime occurs for small q values, where the length scales
probed are larger than the radius of gyration of the aggregate, Rg, and the intensity of
scattering is determined by the suspension as a whole. Conversely, if the wavelength is
small enough (x-rays or neutrons), or if particles are large, information about the surface
Regime Region
Rayleigh 1q> Rg
Guinier 1q≈ Rg
Power Law Rg >>1q>> a
Porod a > 1q
37
of the particles can be obtained in the Porod regime [215]. The power law regime
corresponds to small length scales that fall between the radius of the primary particle, a,
and the radius of gyration of the aggregate. In the power law regime, the intensity of
scattering scales to the power of the fractal dimension such that plotting the intensity D
ownl
oade
d B
y: [K
ansa
s S
tate
Uni
vers
ity L
ibra
ries]
At:
21:4
9 27
Jun
e 20
07 674 C. M. SORENSEN
Figure 22. Schematic representation of the scattered light in-
tensity I (q ) versus q D 4⇤¸¡1sin µ=2, where µ is the scattering
angle, from an ensemble of fractal aggregates of dimension Dwith N monomers per aggregate on a log-log plot. In the en-
semble n is the number density of clusters and Nm is the total
number of monomers, nN D Nm.
in Figure 22. Many examples of optical structure factor measure-
ments on particulate systems exist in the literature (Schaefer et al.
1984; Martin et al. 1986; Hurd and Flower 1988; Zhang et al.
1988; Gangopadhyay et al. 1991; Bonczyk and Hall 1991, 1992;
Dobbins and Megarides 1991; Sorensen et al. 1992b; Koylu and
Faeth 1994a, 1994b; Koylu 1998; Sorensen et al. 1998; Xing
et al. 1999).
An example from some of our early work (Gangopadhyay
et al. 1991) is given in Figure 23 for a soot aerosol in a premixed
methane/oxygen ⇥ame. Increasing height above burner h means
later in the aggregation process. Inspection of Figure 23 shows
that with increasing height, the bend in the optical structure
factor, i.e., where the slope of I versus q goes from 0 to negative,
progresses to smaller q . A cardinal rule is that a change in slope
implies a length scale. In this case the length scale is the overall
aggregate size, and since R � q¡1, this is a direct observation,
albeit qualitative, of the aggregate size increasing with time.
Notice the essentially isotropic scattering at h D 8 nm to indicate
very small particles. Figure 24 presents a more recent example of
scattering from a titania aerosol. Note the scale in q is an order of
magnitude smaller than in Figure 23, hence the clusters of titania
are an order of magnitude larger. In Figure 24 a signi�cant power
law regime is seen with a slope implying D ’ 1:7. Notice also
other general features. The I (q D 0) limit (the Rayleigh regime)
increases with either height above burner or run time. This is
the Tyndall effect (see below) that as a system coarsens as it
scatters more light. On the other hand, the power law regime
Figure 23. Scattered light intensity I (q ) as a function of the
scattering wave vector q for 5 different heights above burner hfor a premixed methane/oxygen ⇥ame.
in Figure 24 is approaching an intensity constant with time as
described above.
Quantitative analysis of the optical structure factor proceeds
in two steps (Gangopadhyay et al. 1991; Sorensen et al. 1992b).
First, the Guinier regime is analyzed to yield the aggregate ra-
dius of gyration. The Guinier equation, Equation (48), may be
Figure 24. Scattered light intensity I (q ) as a function of the
scattering wave vector q for 5 different times (runs) of an ag-
gregating titania aerosol.
Porodregime
Figure 12: Illustration of light scattering regimes when plotting SLS data as intensity vs. scattering vector. Adapted from [196]
versus the scattering vector on a log scale returns a line whose slope is equal to the
fractal dimension:
I(q) ~ qDf (Eq 13.)
38
2.4 Effect of Aggregation on Reactivity
In the case of reactive nanoparticles where reactions occur on the particle surface,
reaction rates will vary, in part, as a function of the concentration of surface in the
suspension. Particle aggregation produces differences in the local particle surface
concentration within an aggregate, and changes in mass transport that depend on the
aggregate morphology (as quantified for example by fractal dimension). It is therefore
not surprising that changes in reactivity have been observed as particles are destabilized
and form aggregates. Gilbert et al. demonstrated that the capacity for metal ion
adsorption by iron oxyhydroxide nanoparticles decreased as fractal dimension increased
[216]. ROS production has been linked to both aggregate size and structure [82, 121, 177,
217]. Lee et al. looked at how the preparation method affects ROS generation by FNPs,
and found that ROS production was likely dependent upon the extent of aggregation
[218]. This difference in ROS generation was hypothesized to result from increased
quenching of the excited triplet state in larger, denser aggregates. Laser flash photolysis
showed that the lifetime of the excited triplet state of individual C60 cages of small
aggregates suspended via surfactants was orders of magnitude greater (tens of
microseconds) than aggregated nC60 clusters (picoseconds) [132].
These observations were supported mathematically in a model proposed by
Hotze et al., who demonstrated the influence of both aggregate size and structure on the
presence of the excited tripled state [219]. Additionally, the Hotze model helps to
39
explain the conflicting ROS production findings for C60 and fullerol in early experiments.
Often suspensions of nC60 were observed to produce little to no ROS while fullerol,
possessing a decreased quantum yield and therefore a theoretically lower production
capacity per cage, produced greater bulk ROS concentrations [127, 132, 220-222]. By
taking into account the morphology of aggregates, the decrease in quantum yield for
fullerol versus fullerene can be compensated for with the smaller, more open aggregates
that are oftentimes formed.
Chae et al., in filtering FNPs showed not only a greater degree of hydroxylation
for smaller aggregates, but also a substantial increase in ROS production as the average
aggregate size decreased [121]. Hu et al. showed that as hydroxylation due to UV
irradiation increased over the course of 15 hours, the degradation rate of furfuryl alcohol
due to singlet oxygen production also increased, attributed to the production of a larger
number of smaller, more stable aggregates [141]. The potential for these transformations,
either of C60 as a result of suspension preparation or of nC60 suspensions resulting from
UV exposure, underscores the need to further investigate the dependency of ROS
production and aggregate stability on surface functionalization.
In studies using TiO2 and ZnO, Jassby et al. observed and modeled the decrease
in �OH generation as particles aggregated in various ionic strengths [177]. In addition to
aggregate size, fractal dimension strongly impacted the reactivity of the material on a
per monomer basis. A previous study by Jassby and Wiesner using coated and uncoated
40
ZnS nanoparticles demonstrated that aggregation to a critical size resulted in the sharp
increase in photoluminescence [217]. It was hypothesized that after sufficient particle-
particle interactions were induced by aggregation, a decrease in surface tension in the
aggregate resulted in greater electron-hole recombination, observed as photon emission
at 284nm. The magnitude of the photoluminescence is likely due to aggregate structure,
with denser aggregates absorbing less incident light and producing less
photoluminescence, though the exact interplay is not yet well determined [29].
2.5 NPs in Environmental Waters
Nanoparticles released into environmental waters will be exposed to various
chemistries that will affect their aggregation state and ROS production capacity. These
factors influencing NP aggregation include inorganic ions, natural organic matter,
metals and organic pollutants. Some ions (e.g. phosphate, nitrate) and organic pollutants
(e.g. pesticides, PCBs) are often found at elevated concentrations in urban waters [223].
While there is significant variation, typical surface waters range from pH values
of 5.7 to 9, with carbonate, chloride, sulfate, and nitrate the most common anions at
approximately µM concentrations [182, 224]. Carbonate is a particularly efficient radical
quencher, and although the resultant carbonate radical is a weaker oxidant than �OH, it
can itself participate in a number of reactions [225].
In addition to promoting aggregation due to compression of the electrical double
layer, specific interactions with the TiO2 surface will likely further impact stability and
41
ROS generation [29, 177, 226]. Phosphorus concentrations are typically low in natural
freshwaters, though the extensive use of fertilizers in both urban and agricultural
landscapes creates nearly ubiquitously elevated concentrations in these locations.
Additionally, phosphate is known to specifically interact with the TiO2 surface, which
has been seen to increase NP stability [227, 228]. As NPs are often negatively charged,
the presence of metal cations, such as sodium, magnesium, calcium and aluminum, will
play a significant role in inducing aggregation.
Natural organic matter (NOM) is a ubiquitous chromaphore, primarily
comprised of degradation products of biomolecules existing in surface waters at
concentrations widely varying around 1 mg L-1 (ranging from 0.1 to 100, depending on
water type) [182, 229, 230]. NOM coatings such as humic and fulvic acids have been
shown to increase nanoparticle stability and transport by increasing steric hindrance
[184, 185, 231]. Furthermore, the redox chemistry and photosensitizing ability of NOM
will likely complicate attempts to determine net photoreactivity in mixed NP/NOM
systems [65-68].
Under environmental release scenarios a nanoparticle or aggregate will interact
with suspended background particles, producing mixed aggregates [232-235]. The vast
difference in particle number concentrations makes it likely that aggregates will consist
mostly of colloidal matter with nanoparticles incorporated into the structure. What
remains unclear, however, is how this combination will impact NP reactivity. In some
42
instances, the creation of heteroäggregates has been shown to increase ROS production.
This has been demonstrated for mixtures of TiO2 and SiO2 processed at high
temperatures [236-238], though nanoparticles released into surface waters and exposed
to UV light are unlikely to be exposed to such extreme conditions. Conversely, the same
effects of aggregate size and structure seen in homoäggregates to impact NP reactivity
will take place during heteroäggregation. While some studies have looked at
heteroäggregation under environmentally relevant conditions, very little research into
the impacts of these interactions on NP reactivity has been undertaken [170].
2.6 Possible Risks Posed by Nanomaterials
While the working definition of NPs hinges on one dimension being less than
100nm, the useful properties (e.g. increased reactivity, dissolution, toxicity) resulting
from this small scale are the primary drivers of NP production and innovation rather
than their bulk counterparts [107, 226, 239, 240]. However, along with the promise of
emerging technologies come concerns surrounding the impact of inevitable NP release
into the environment. The opportunity exists with NPs to enable advancements of
technology in an environmentally responsible manner, and from the onset, a concerted
effort has been made to concurrently develop the technology and understand the risk
associated with the implementation of said technology [7]. This drive has come from
government regulation as well as public concern, and at the heart of these concerns are
still-unanswered questions about how the intriguing properties of a NP that we look to
43
harness, such as reactivity, are affected by the fundamental interactions (e.g., transport,
aggregation, sorption) that will occur when exposed to various environments [10, 18,
241, 242].
Assessing the risk of engineered NPs in a timely fashion so as not to be outpaced
by, or to impede, technological advancement has proven difficult. While the challenges
presented by NPs are not dissimilar to other emerging contaminants, conventional
chemical toxicology does a poor job at predicting risk for NPs, as actual exposure often
is not the same as predicted exposure [242]. Environmental impacts of NP release have
not yet been observed or established, however monitoring and research are unable to
keep pace with the increasing development of nanotechnology [7]. The myriad potential
exposure conditions, endpoints and environmental systems to be considered for even a
single NP render a comprehensive understanding of the associated risk an intractable
problem [243]. Neither the time, the funding, nor the research hours are available to
realize every potential permutation in which a NP could impact the environment.
Furthermore, the analytical instrumentation to do so is often inadequate. There is
general agreement that research should be undertaken at realistic exposures, though in
the absence of observed impacts, a consensus of what constitutes “environmental
relevance” has not yet been reached [242].
Progress has been made in toxicity and fate/transport modeling based on the
combined data of individual experiments but to date this has been slow, and the vast
44
number of possible combinations makes studying each condition individually untenable
[244]. To effectively utilize the models we have and are developing, the most likely
scenarios should be given priority. Data generated from these targeted tests will then
become model inputs to more effectively improve predictability [242]. The emerging
nanoinformatics field looks to facilitate both the identification of what constitutes
relevant data as well as the integration of these data into workable databases [241].
2.6.1 Need for Standardized Systems and Tests – Functional Assays
Predicting the impact of NP photoreactivity relies on the idea that
physicochemical properties of NPs determine exposure and hazard (risk), yet
approaching NPs this way has not lead to effective guidelines or regulations. This is
largely due to the path-dependent changes to these physicochemical properties that
govern transport, reactivity, etc. As discussed above, released NPs will significantly
transform and age, resulting in different behaviors and reactivity depending on the
environments encountered [29, 245]. NP form, shape, aggregation state, surface area,
and coatings will all be impacted by environmental conditions, and these will in turn
dictate effective dose, toxicity, bioavailability and reactivity. Tracking, measuring, and
predicting these changes are not trivial tasks as these interactions are not yet well
understood. Perhaps because of this, with no established ageing protocols or agreed
upon realistic endpoints, most research has focused on pristine or minimally aged NPs
[242]. Identifying the major pathways through which NPs might reasonably travel and
45
impact their environment would lead to risk assessments for a focused list of standard
chemistries and release scenarios - a much more attainable goal [10, 11].
To facilitate this, it has been proposed to establish an agreed upon set of
representative environmental conditions that can be paired with specifically monitored
endpoints for which a NP should reasonably be investigated [241, 242]. The use of
functional assays (FAs) can yield data at specific endpoints for NPs in complex reference
systems that take into account this path-dependent nature. Thus, under representative
environmental conditions, FAs will employ behavioral tests (akin to BOD5 or KOW
measurements) to measure an expected outcome or impact of the NP, endpoints that
will incorporate but not require a fully mechanistic understanding of NP behavior.
Ultimately, FAs will provide a framework for individualized experimental
measurements to more effectively test and describe scientific theory, determining rates
of processes or interactions that will enable parameterization of models [241]. In the
absence of agreed upon representative environmental conditions, investigating trends in
nanoparticle reactivity and the primary variables that impact ROS production represents
an important step forward in our understanding.
46
3. Methods Nanoparticle reactivity will depend greatly on aggregation state and solution
chemistry. Therefore, to effectively study the impact of surface chemistry, not only do
reactive endpoints (e.g. ROS generation, bacteria inactivation) need to be measured, but
NP suspensions should also be well characterized (e.g. aggregate size and structure,
surface charge). This chapter provides common methods for the NP photoreactivity
studies in this dissertation. Specifics to the three cases considered in this research are
included in Chapter 4 (Sections 4.1.2, 4.2.2 and 4.3.2).
3.1 NP Suspension Preparation
The method of preparing consistent, stable suspensions of NPs depends largely
on the type of NP being used and the solution into which it is to be dispersed. For NPs
that do not readily disperse in solution, probe sonication was performed using a high-
energy probe (S-4000, Misonix, Qsonica, LLC, Newtown, CT) with a 1/2” diameter tip.
3.1.1 FNP Suspensions
C60, C60(OH)6, C60(OH)24 and C60(NH2)6 were purchased in powder form from
MER Corp. (Tucson, AZ). A colloidal suspension of fullerene (aqu-nC60) in NanoPure
water (NPW) was prepared by probe sonication of 100 mg C60 powder in 200 mL NPW
(resistivity > 18 MΩ cm and TOC< 30 µg L-1) for 10 hours operated in pulse mode (10
min ON, 20 min OFF) in the dark. Following sonication, the suspension was filtered
through a 100 nm nylon membrane filter (Magna, GE Osmonics, Minnetonka, MN). The
47
filtrate was then concentrated via rotavaporator (Büchi, Flawil, Switzerland). DLS
measurements of the suspension before and after concentration indicated no change in
average aggregate size.
Well dispersed stock suspensions of C60(OH)6, C60(OH)24, and C60(NH2)6 were
prepared by adding powder directly to ultrapure water and stirring overnight and
filtering to < 100 nm. The total carbon concentration of all stock suspensions post
filtration was measured (TOC-L, Shimadzu, Japan), and stocks were diluted to a
concentration of 5 mg C L-1 for experiments.
3.1.2 TiO2 Suspensions
Suspensions of TiO2 were prepared through probe sonication of NP power in the
desired media. 2 mg of powdered TiO2 NPs (P25 Aeroxide TiO2, Evonik Industries,
Essen, Germany) were weighed with an analytical microbalance (Mettler Toledo, USA)
in the hood and transferred to 30 mL of NPW. To this 20 mL was added to reach a final
volume of 50 mL, resulting in 40 mg L-1 TiO2 in a solution of desired chemistry.
Following the protocol of Taurozzi et al. NP suspensions were placed in an ice bath and
probe sonicated at 60% amplitude in pulse mode (12 s on 3 s off) for a total of 5 min
sonication time, resulting in a suspension having a median aggregate radius of
approximately 100 nm as measured by DLS.
TiO2 suspensions close to the PZC (6.2-6.3) will not remain in suspension for
appreciable lengths of time, and aggregates will double in size in tens of minutes. For
48
this reason, pH values were adjusted sufficiently far from the PZC, and the use of a pH
buffer (e.g. phosphate) was frequently employed. Even within a buffered solution,
overnight storage of suspensions is not recommended, and new TiO2 stocks were
created immediately prior to each use.
3.2 UV Illumination
UV irradiation was performed in a Pelco UVC2 Cryo Chamber (Ted Pella, USA)
equipped at the top with a manifold containing two Phillips TL-D 15W BLB bulbs
(Phillips, Netherlands) The output of these bulbs centered at 365 nm and was modeled
as a normal distribution with mean at 365 nm and standard deviation of 10 nm. The
lamps were separated from the chamber by a shutter, enabling the flux of light to be
interrupted and samples collected without turning off the lamps. Prior to experiments,
the lamps were turned on and allowed to stabilize, as an initial spike in UV flux
occurred which then decayed to a stable output after approximately 30 minutes.
Temperature was held constant at 22 ± 2 °C though the use of a chiller. Irradiance was
measured throughout the duration of experiments with a radiometer equipped with a
UV-A filter (UVX , UVP, Inc., USA and ILT1400 and SEL033 UV-A filter, International
Light Technologies, USA).
The flux of UV light was observed to depend on location in the UV chamber,
both along the z-axis (distance from the bulbs) as well as x-y positioning. Four locations
of equal flux were identified such that measurements could be run in triplicate and the
49
radiometer employed at the same time. Magnetic stir plates were positioned at these
locations, and experiments were run as stirred batch reactors. The geometry of the
reaction beaker (H=3.5 cm, Ø=5 cm) produced a 0.5 cm thickness liquid layer, including
the stir bar (3 x 10 mm). For experiments in which volatilization was a concern, reactors
were covered with glass petri dishes, though this reduced the flux slightly.
3.3 ROS Measurements
ROS production was measured with fluorescent probes and quenching agents as
described below. For 1O2 and �OH, standard curves were produced to quantify ROS
generation.
3.3.1 1O2 Measurements with SOSG
1O2 concentrations were measured as the increase in fluorescence units (FSU) as
Singlet Oxygen Sensor Green (SOSG, Molecular Probes-Invitrogen) was oxidized to
SOSG-EP. Excitation of SOSG-EP between 460-490 nm results in fluorescence between
510-550 nm. A Modulus fluorometer (Modulus Single Tube 9200, Promega, USA)
equipped with a blue optical kit filter was used to measure fluorescence. The blue
optical filter consisted of an excitation wavelength at 460 nm and a band pass emission
filter ranging from 515 – 570 nm, which corresponds well with SOSG-EP.
SOSG (Invitrogen, Thermo Fisher Scientific, USA) was ordered in packaged
aliquots of 100 µg that were stored at -20 ˚C. The probe was prepared by suspending one
aliquot of SOSG in 33 µL of methanol. This was combined with 0.97 mL of NPW by
50
carefully drawing the solution in and out of a pipette tip, resulting in a SOSG stock
concentration of 165 µM. 0.1 mL of this solution was spiked into 10 mL samples to
measure singlet oxygen at a final SOSG concentration of 1.65 µM. β-carotene was used
as a quenching agent at a final concentration of 50 µm. Stocks in methanol stored in the
freezer were diluted to a 1 mM solution and bath sonicated to ensure dissolution prior to
use.
3.3.1.1 Standard Curve Preparation
Singlet oxygen production was determined by comparing the measured
fluorescence of SOSG in fullerene suspensions versus a Rose Bengal (RB) standard
curve. Oxidation of SOSG at various RB concentrations was measured over 5 minutes,
and a linear fluorescence response was determined for a RB concentration range of 0.005
– 0.1 µM.
The rate of photons hitting the sample per second was calculated from the output
of the lamps, the liquid surface area, and the Plank-Einstein equation as follows:
rphoton =irr ⋅a ⋅λh ⋅c
(Eq 14.)
where irr is the irradiance in W/cm2, a is the illuminated area of the reaction vial, c is the
speed of light, λ is the wavelength, and h is plank’s constant. Combining this with
absorption measurements of the RB standard solution at 365 nm, ABS365, the quantum
yield, Φ, and Avogadro’s number, NA, gives the rate of singlet oxygen production in
molar concentration per time:
51
r1O2 =rphoton (1−10
−ABS365 )φV ⋅NA
(Eq 15.)
Knowing this, the fluorescence response of SOSG can then be equated to the
singlet oxygen production rate as follows. Accumulation of singlet oxygen in suspension
can be modeled as
d[1O2 ]dt
= r1O2 − kd[1O2 ] (Eq 16.)
where kd is the rate of quenching in water (kd = 2.4x105 sec-1) [373]. Making the steady
state approximation, an equation for [1O2]ss, in moles, can be calculated as
[1O2 ]ss =r1O2kd
(Eq 17.)
Thus, from the Rose Bengal standard suspensions, the fluorescence response due
to a given steady state singlet oxygen concentrations can be calculated as
dFSUdt
= k1O2 [1O2 ]ss[SOSG] (Eq 18.)
where FSU is the detected fluorescence resulting from the cumulative reaction of SOSG
with the produced 1O2 over the irradiance time t, k1O2 is the second order rate constant,
and [1O2]ss is the steady state 1O2 concentration. By using a sufficiently large [SOSG]
compared to produced 1O2 and short irradiance times, the concentration of unreacted
SOSG in solution can be assumed constant, giving the pseudo first order kinetics:
52
dFSUdt
= k '1O2 [1O2 ]ss (Eq 19.)
Here, k’102 is the pseudo-first order rate constant whose units are FSU uM-1 min-1.
Typically, pseudo-first order and first order rate constants have units of concentration
per concentration per time, which reduces to simply units of per time. Because
fluorescence, which does not have units of concentration, was modeled, the reactivity
rate constant is slightly modified. The observed increase in fluorescence was plotted
versus calculated singlet oxygen concentration, and the slope (with units of FSU uM-1
min-1) was used to convert the observed rate of fluorescence into the singlet oxygen
concentration in solution. Concentrations of irradiated fullerene suspensions were
adjusted to maintain fluorescent responses within the linear standard range.
3.3.2 �OH Measurements with TA
Hydroxyl radical production tests were performed on samples of 20 mg L-1 P25
TiO2 through the hydroxylation of TA to 2-HTA (ex 315, em 425). 0.5 mM TA stock
solutions were prepared by adding 8.36 g TA to 100 mL NPW. The solubility of the
probe increases with pH so a buffer solution or pH adjustment was necessary for
preparation. Stock solutions were stirred on a magnetic stir plate for 24 hours to ensure
dissolution, filtered with a disposable sterile 0.22µm cellulose acetate filter system
(Corning, USA) and stored in the dark. TA was dosed at a final concentration of 125 µM
in solution.
53
Hydroxyl radical measurements were made by adding a 0.5 mL sample aliquot
to 0.75 mL deionized water in a 2 mL centrifuge tube. For conditions of low expected
�OH production, this dilution step was skipped, and 1.5 mL was sampled directly. The
presence of NPs was observed to influence fluorescence measurements due to light
scattering, so samples were centrifuged for 5 minutes at 12,000 RPM to separate the TiO2
from the solution. 1 mL of the resulting supernatant was transferred to a plastic cuvette
(10 mm optical path) and fluorescence was measured (Varian Eclipse fluorometer,
Agilent, USA) at an excitation wavelength set to 315 nm. 2-HTA fluorescence was
recorded from 325-540 nm to collect the full emission peak. The gain on the
photomultiplier was adjusted depending on experiment to ensure that fluorescence from
small amounts of ROS generation was detectable but did not saturate the detector at
longer timepoints. All samples were measured using the same gain setting to enable
comparison between experiments. No auto-fluorescence was observed in blanks or dark
samples, nor did overhead fluorescent lighting influence results. N-acetyl-L-cysteine (N-
AC) (Sigma-Aldrich) was used as a �OH quencher. 16.3 mg N-AC was dissolved in 10
mL NPW and filtered in a 0.22 µm syringe filter to sterilze, resulting in a 10 mM stock
concentration. For quenching experiments, 0.2 mL was spiked into 10 mL samples for a
final concentration of 200 µM.
54
3.3.2.1 Standard Curve Preparation
Fluorescence was converted to 2-HTA concentration using a standard curve
created by diluting a 0.125 mM 2-HTA stock solution to 6 points ranging from 0.0025 to
0.1 M, as determined by fluorescence limits and detector saturation. Hydroxyl radical
concentrations were estimated with the following reaction, assuming 80% trapping
efficiency for OH• by TA [87] and considering sample dilution prior to centrifugation:
TA+ •OH→ 2−HTA (Eq 20.)
3.3.3 O2�- Measurements with HE
Superoxide (O2�-) was detected as the increase in fluorescence from 540 to 800
nm (max 586 nm, excitation at 510 nm) resulting from the transformation of
dihydroethidium (DHE) (Invitrogen, Thermo Fisher Scientific, USA) in solution to 2-
hydroxyethidium (2-HE). Stock solutions of HE were prepared by dissolving 1 mg HE in
100 µL DMSO to which was added 900 µL NPW, resulting in a 3.17 mM stock solution.
For quenching experiments, 160 µL were spiked into 10 mL samples for a final
concentration on 50.7 µM (0.16% v/v DMSO).
Superoxide dismutase (SOD) (Sigma-Aldrich) was used as the O2�- quenching
agent. Stock solutions were made at 3,000 U mL-1 by adding 10 mL NPW to the 5.8 mg
container and stored at -20 ˚C. For quenching experiments 100 µL SOD stock was spiked
into 10 mL samples for a final concentration of 30 U mL-1.
55
3.4 Additional Reactivity Endpoints
When employing biological endpoints, all glassware, stir bars and pipette tips
were autoclaved prior to use. With the exception of UV exposure experiments, reactions
were covered in aluminum foil to limit contamination. Stocks and samples were kept on
ice during experiments and during the transfer of samples.
3.4.1 Viral Inactivation
For viral inactivation experiments, a 0.1x phosphate buffered saline (PBS)
solution was used to maintain virus viability. Bacteriophage MS2 (American Type
Culture Collection (ATCC) 15597-B1) was propagated on E. coli (ATCC 15597) and
enumerated using the double-top agar assay. A highly purified, high-titer virus stock in
PBS was prepared as follows. First, MS2 was cultured on double-layer agar plate with E.
coli as the host incubated at 37 °C. Next, the phages on the top agar were extracted into
1x PBS and precipitated using 10% polyethylene glycol at 4 °C for 1 h. The precipitated
phages were centrifuged at 10000g for 15 min at 4 °C, and the supernatant was decanted
and the pellet resuspended in PBS and incubated overnight at 4 °C. The residual
bacterial debris and polyethylene glycol in the phage suspension were removed by
centrifugation at 5000 g for 10 minutes at 4 °C and the addition of chloroform (1:1
sample-to-chloroform), respectively. The phage-containing aqueous phase was
separated from the organic phase by centrifugation as before. The suspension was
furthermore concentrated by ultracentrifugation at 103,000 g for 2.5 h at 12 °C, and the
56
phage pellet was then resuspended in PBS. The viruses were counted in terms of plaque
forming units (PFU). This procedure yielded an initial stock concentration of 1.1x1012
PFU mL-1. MS2 toxicity was calculated as log inactivation relative to time 0 min.
3.4.2 Bacterial Inactivation
All bacterial inactivation experiments employed a Minimal Davis media (MD)
designed to stabilize the particles, buffer the pH at 7.5, and maintain bacterial viability,
as described by Lyon et al. [366]. Briefly, the recipe for 1x MD is a follows: 45 mg
K2HPO4, 50 mg (NH4)2SO4, 25 mg sodium citrate, and 5 mg MgSO4�7H2O were dissolved
in 100 mL NPW. MD stock solutions were created at a 10x concentration and bath
sonicated, after which they were sterilized via filtration with at 0.22 µm cellulose acetate
disposable filtration unit, and stored in the fridge until use. The cavitation induced by
probe sonication was observed to sufficiently sterilize TiO2 stock suspensions.
Pure cultures Gram Positive Bacillus subtilis strain 168 ATCC 23857 and Gram
Negative Acinetobacter baumannii ATCC 49466 (ATCC, Manassas, VA, USA) were grown
overnight at 37 oC in tryptic soy broth. Cells were pelleted at 8,000 rpm for 10 minutes
and washed thrice with MD. Cells were then resuspended in MD to a final concentration
of 109 cells mL-1. 2 mL aliquots of the bacteria suspension were added to each batch
reactor.
Samples taken from the reactors for inactivation measurements were serially
diluted to obtain a final concentration of 103 CFU mL-1. 100 μL of the final dilution was
57
then spread onto tryptic soy agar plates in triplicate and allowed to grow overnight at 37
oC. The resulting colonies were then counted, and ROS and CPF toxicity was calculated
as log inactivation relative to time zero.
3.4.3 Inactivation Kinetics and ROS Dose
Inactivation kinetics were calculated using the Chick-Watson model employing
the idea of CT, the dose measurement for conventional disinfectants, where
CT = FSU(t)k 'ROS
= [ROS]0 t (Eq 21.)
Here, CT is the disinfectant concentration, C, multiplied by the contact time, t. The
disinfectant concentration, C, is the steady state ROS concentration probed on or very
near the surface of the NP, calculated using the experimental standard curve, above. Log
inactivation was then plotted vs. CT to determine dose response.
3.4.4 CPF Degradation
Chlorpyrifos (99.5%, Chem Service, West Chester, PA, USA) stocks were made in
methanol at 1.5 g L-1 and stored in the fridge for up to 1 month. Working stocks were
created at 1.5 mg L-1 by spiking 10 µL of the stock in 90 mL NanoPure water on a stir
plate. To this, 10 mL of 10x MD media was added dropwise. For all experiments, CPF
was spiked into samples at 375 µg L-1. Stocks of CPF degradation products CPF Oxon
(analytical standard, AccuStandard, New Haven, CT, USA) and 3,5,6-trichloro-2-
pyridinol (TCP) (analytical standard, Sigma-Aldrich) suspensions were similarly
58
prepared. Chlorpyrifos, CPF Oxon, and TCP concentrations were determined via LC/MS
as outlined in Chapter 4.3.2.
3.5 Aggregate Characterization
3.5.1 Transmission Electron Microscopy
High-magnification images of suspensions were obtained by transmission
electron microscopy (TEM) (FEI Tecnai G2 Twin, Hillsboro, OR). For NP imaging, 10 µL
was deposited on a lacey carbon/Cu grid (300 mesh, Electron Microscopy Sciences, USA)
and dried in air before TEM measurement. MS2 bacteriophages were negatively stained
with 2.5% phosphotungstic acid (Fluka, Milwaukee, WI) prior to imaging. Bacteria were
deposited on the grid, fixed with 2.5% glutaraldehyde, and sequentially dehydrated
with graded ethanol and washed with 1x PBS in order to maintain structural integrity.
3.5.2 Aggregate Size by Laser Diffraction
Laser diffraction (Mastersizer 3000, Malvern Instruments, England) was used to
determine TiO2 aggregate size and fractal dimension (Section 3.5.4). In-line light
scattering measurements were performed using a peristaltic pump located downstream
of the instrument by which samples were fed from a beaker continuously mixed on a
magnetic stirrer at 400 RPM. The pump rate was set at approximately 15 mL min-1 to
minimize shear within the tubing, and 600 mL samples were used so that the suspension
was not recirculated, avoiding changes in aggregation state due to the pump. Prior to
data collection, background measurements were collected using the same solution
59
chemistry and sample flow rate as in NP aggregation samples. Time resolved size
measurements, reported in terms of median hydrodynamic diameter (D50) both as
number weighted and volume weighted intensity, were set to 10 s duration with a 100 s
delay, resulting in measurements being collected every two minutes.
3.5.3 Particle Size by Dynamic Light Scattering
Dynamic light scattering (DLS) measurements were performed with the
Zetasizer Nano ZS (Malvern Instruments, England), which is equipped with a He-Ne
laser (633nm) and a detector fixed at 173°. A minimum of 3 measurements were
collected, with aggregate size reported both as unweighted and number weighted. The
polydispersity index (PDI) is a value ranging from 0 – 1 and indicates the degree to
which aggregates in suspension are uniformly distributed with larger PDI values
indicating more heterogeneous aggregate sizes.
3.5.4 Fractal Dimension by Static Light Scattering
Static light scattering (SLS) measurements were collected with the Mastersizer
and were performed concurrent with aggregate size measurements. Data recorded for
each sample by an array of detectors consisted of the scattering intensity, I(q), over a
range of scattering vectors, q, corresponding to 0.8 - 42°. The aggregate fractal
dimension, Df, was determined as the negative slope of the linear power law region
when plotting log(I(q)) vs. log(q) [215]:
60
I(q) = q−Df (Eq 22.)
The multiple detectors of the Mastersizer correspond to scattering vectors from
0.18 – 9.47 µm-1, with q related to the detector angle, theta, as follows
q = 4πnλsinθ2
(Eq 23.)
where n is the refractive index of the medium (1.33 for water) and λ is the wavelength of
scattered light (633 nm) [194, 215]. Equation 22 holds when
qRh >1 (Eq 24.)
i.e. for length scales, (q-1) less than the radius of the aggregate, Rh.
At each timepoint, Df was calculated via linear regression over the power law
region of the scattering curve. Linearity was determined to hold when the difference
between the local slope for a given q value and the slope of a line drawn between the
same q value and the final q value was no greater than 12%. The process of calculating
fractal dimension from scattering data was automated by code written in Matlab.
3.5.5 Electrophoretic Mobility and pH Titrations
Particle electrophoretic mobility (EPM) was measured a minimum of 3 times
(Zetasizer Nano ZS) and reported as zeta potential (calculated via the Henry equation).
As EPM is highly dependent on solution chemistry, measurements were performed
under the same conditions as those being tested.
61
TiO2 nanoparticle pH titrations were performed on 50 mL suspensions of 20 mg
L-1 P25 TiO2 at various anion concentrations. Suspensions were continually stirred in a
beaker placed on a magnetic stir plate while pH was monitored and adjusted with either
0.01 M or 0.1 M KOH or HCl. At each pH step, the rate of stirring was increased and
sufficient acid or base was added to move the solution pH approximately 0.2 units. This
was allowed to mix briefly before stirring was reduced again and pH was allowed to
equilibrate. 0.7 mL sacrificial aliquots were taken for electrophoretic mobility
measurements and the isoelectric point (IEP) was identified as the pH for which the
EPM of a suspension was zero.
62
4. Results and Discussion
4.1 Bacteriophage Inactivation by UV-A Illuminated Fullerenes: Role of Nanoparticle-Virus Association and Biological Targets
Abstract
Inactivation rates of the MS2 bacteriophage and 1O2 generation rates by four
different photosensitized aqueous FNP suspensions were observed to follow the same
order: aqu-nC60 < C60(OH)6 ≈ C60(OH)24 < C60(NH2)6. Alterations to capsid protein
secondary structures and protein oxidation were inferred by detecting changes in
infrared vibrational frequencies and carbonyl groups respectively. MS2 inactivation
appears to be the result of loss of capsid structural integrity (localized deformation) and
the reduced ability to eject genomic RNA into its bacterial host. Evidence is also
presented for possible capsid rupture in MS2 exposed to UV-A illuminated C60(NH2)6
through TEM imagery and detection of RNA infrared fingerprints in ATR-FTIR spectra.
FNP-virus mixtures were also directly visualized in the aqueous phase using a novel
enhanced dark-field transmission optical microscope fitted with a hyperspectral imaging
(HSI) spectrometer. Perturbations in intermolecular extended chains, HSI, and
electrostatic interactions suggest that inactivation is a function of the relative proximity
between nanoparticles and viruses and 1O2 generation rate. MS2 log survival ratios were
linearly related to CT (product of 1O2 concentration, C, and exposure time, T)
demonstrating the applicability of Chick-Watson kinetics for all fullerenes employed in
63
this study. Results suggest that antiviral properties of FNPs can be increased by
adjusting the type of surface functionalization and extent of cage derivatization thereby
increasing the 1O2 generation rate and facilitating closer association with biological
targets.
4.1.1 Introduction
Generation of singlet oxygen (1O2) by photosensitized FNPs is known to
inactivate bacteria and viruses, which is of interest to water disinfection [246-248],
photodynamic therapy [249], and other medical and environmental applications [250].
Ecological risks arising from unplanned releases of nanoparticles can also be partially
attributed to reactive oxygen species (ROS) production or oxidative stress [81, 251, 252].
Therefore, toxicological effects of fullerenes can be expected to be strongly dependent on
the effective delivery of exogenous ROS generated by nanoparticles to the biological
target. It is known that ROS production is determined by the morphology of
nanoparticle aggregates, types of surface functional groups, and extent of surface
derivatization [253-255], all of which also influence the effective separation distance
between the nanoparticles and their biological targets. In this context, we hypothesize
that both the 1O2 generation rate and the relative proximity between fullerenes and
viruses are important determinants of inactivation particularly since the lifetime of 1O2
in aqueous solutions is only 3 – 4 ms [256, 257]. Limited data available in the literature
supports our assertion even though direct evidence has not yet been provided. For
64
example, the differential inactivation of MS2 and Escherichia coli, by illuminated TiO2
nanoparticles was likely correlated to varying amounts of association with TiO2 [258].
Additionally, malonic acid adduct fullerenes were more phototoxic to Jurkat cells
compared with a dendritic fullerene because they were postulated to associate more
closely with the biological target even though their 1O2 quantum yield was lower [259].
Also, electrostatic attraction between negatively charged viruses and cationic aminated
FNPs very efficiently inactivated MS2 [33]. Even close association with natural organic
matter (NOM) enhances MS2 inactivation since 1O2 concentrations are elevated within
the macromolecule [251]. While these studies indirectly postulate the significance of
nanoparticle-microorganism association, the simultaneous interplay of proximity and
1O2 generation during virus inactivation has not been explicitly examined to date.
While ROS-mediated antiviral activity of fullerenes and their derivatives has
been amply demonstrated e.g. [33, 121, 246, 260], much less is known about the
underlying virucidal mechanisms. Recent evidence points to oxidation of capsid
proteins by exogenous ROS [247, 261, 262], but more detailed information on changes in
structure is necessary to better understand underlying causes of inactivation. We
hypothesize that 1O2 modifies conformation of virus capsid proteins leading to loss of
infectivity. The F-specific ssRNA coliphage MS2 was employed as a surrogate for
nonenveloped enteric human viruses and to facilitate comparisons with other studies
e.g. [33, 121, 246, 262].
65
Herein, we test the hypotheses that (i) virus inactivation is caused by changes in
capsid integrity arising from oxidative protein conformational alterations or
deformation and (ii) the rate of virus inactivation by photosensitized FNPs is controlled
by their relative proximity as well as the 1O2 generation rates. MS2 inactivation kinetics
and 1O2 generation rates were measured in the presence of UV-A illuminated
suspensions of aqu-nC60, C60(OH)6, C60(OH)24, and C60(NH2)6. Oxidative damages to MS2
capsid proteins were characterized using attenuated total reflectance Fourier transform
infrared spectroscopy (ATR–FTIR), OxyBlot protein oxidation assay, and sodium
dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). A novel hyperspectral
microscopy technique was utilized to directly visualize association between
nanoparticles and viruses in the aqueous phase. Hyperspectral imaging (HSI) allowed
the superposition of a large number of images (~ 400) at high resolution (~ 1.5 nm) and
identifying their components by matching their spectral signature instead of material
color eliminating the need for sample dehydration. While HSI has been employed for
military applications, mineral exploration, quality assurance, food safety, forensics, and
health care e.g. [263], this is one of the first reports of its use in virus/nanoparticle
identification. Further, the applicability of Chick-Watson disinfection kinetics “CT”
(product of 1O2 concentration C and exposure time T) for ROS-based inactivation is
discussed with particular emphasis on the relative proximity of nanoparticles to the viral
target and physicochemical parameters associated with 1O2 generation.
66
4.1.2 Materials and Methods
4.1.2.1 Cultures, Phage Purification and Enumeration
Bacteriophage MS2 (American Type Culture Collection (ATCC) 15597-B1) was
propagated on E. coli (ATCC 15597) and enumerated using the double-top agar assay. A
highly purified, high-titer virus stock in PBS (1013 PFU (plaque forming unit) mL-1), free
of bacterial debris and media residues was prepared by centrifugation, precipitation,
extraction, and ultracentrifugation [121].
4.1.2.2 Fullerene Suspensions
C60 (99.9%), C60(OH)6, C60(OH)24 and C60(NH2)6 were purchased in powder form
(MER Corp., Tucson, AZ). The aqueous fullerene suspension (aqu-nC60) was prepared
by ultrasonication and filtration (see Appendix A for details). Well dispersed stock
suspensions of C60(OH)6, C60(OH)24, and C60(NH2)6 were prepared by adding powder
directly to ultrapure water and stirring overnight and filtering to < 100 nm. Particle size
distributions were obtained via dynamic light scattering at 90° (ALV/LSE-5004, Langen,
Germany) (see Appendix A, Figure 35). Electrophoretic mobilities were measured via
Zetasizer Nano (Malvern Instruments, Bedford, MA). A concentration of 5 mg C L-1
(measured as total carbon, TOC-5050A, Shimadzu) was employed for all nanoparticles.
4.1.2.3 Irradiation and 1O2 Detection
1O2 concentrations were measured as the increase in fluorescence units (FSU)
(Modulus Single Tube 9200, Turner Biosystems) of Singlet Oxygen Sensor Green (SOSG,
67
Molecular Probes-Invitrogen) and quantified through Rose Bengal standard curves
(Appendix A, Figure 36 and Figure 37). Two 15 W fluorescent bulbs (Philips TLD
15W/08) provided low-pressure UV irradiation, which had an output spectrum peak at
365 nm in the UV-A region and an intensity of 2.3 mW cm-2 as measured by a UVX
Radiometer (UVP, Inc. Upland, CA) (see Appendix A for output spectrum, Figure 38).
Test suspensions consisted of (1) MS2 + UV-A, (2) MS2 + FNP + UV-A, (3) MS2 + FNP in
the dark, and (4) MS2 + FNP + β-carotene + UV-A. All vials were continually stirred,
experiments were performed in triplicate, and β-carotene, an efficient 1O2 scavenger
served as a negative control.
4.1.2.4 ATR-FTIR Spectroscopy
The mid-infrared spectra (1800-1000 cm-1) of MS2 (in solid form) treated with
UV-A alone and UV-A sensitized fullerenes were collected using a Nicolet Magna 750
spectrometer equipped with infrared source, DTGS detector, and a KBr beamsplitter.
Samples were prepared by first freezing a 45 mL sample in polypropylene vials
overnight at -81 °C. Next, the vials were transferred to a sample holder attached to a
freeze drier operating under vacuum for 48 h. The sample pellet from each vial was
harvested and stored in 1.5 mL centrifuge tubes at 4 °C. Sufficient sample powder was
deposited to cover the entire Ge internal reflection element (IRE) surface and the IR
spectrum for each sample was obtained by averaging 512 scans collected at 4 cm-1
resolution. Background spectra were collected on a clean Ge IRE. All spectra were
68
processed using OMNIC v.5.29 software (Nicolet-Thermo Electron). Absorption peaks
were assigned based on the second derivative of the original spectra and the sample
constituents contributions to the peaks were described based on amino acids, proteins,
and viruses [264, 265]. The peaks in the region 1700-1500 cm–1 were decomposed to
separate overlapping peaks and then fitted to a Lorentzian-Gaussian model.
4.1.2.5 OxyBlot Protein Oxidation Assay
Prior to quantifying the carbonyl groups formed by the reactions of 1O2 with
amino-acid side chains, MS2 samples after 5 minutes of irradiation were concentrated to
get 3-4 µg µL-1 of total proteins (measured by Modified Lowry Assay). 10 µL aliquot of
denatured-proteins was reacted with 2,4-dinitrophenylhydrazine (DNPH) for 15 min to
form 2,4-dinitrophenylhydrozone (DNP-hydrazone), which confirms the presence of
carbonyls in the sample. The amount DNP-hydrozone product formed was quantified
by following a series of steps involving SDS-polyacrylamide gel electrophoresis (SDS-
PAGE), Western blot analysis and treatment with antibodies, and chemiluminescent
reagents (see Appendix A for detailed explanation). The oxidative status of each protein
was then analyzed quantitatively by comparing the signal intensity (at 428 nm) of the
same protein in different lanes on the same gel. A negative control was also prepared as
described but without protein.
69
4.1.2.6 SDS-PAGE
SDS-PAGE was used to examine MS2 phage protein samples for the presence of
cross-linkages induced by 1O2 during illumination of fullerene suspensions [247].
4.1.2.7 Dark-Field based Hyperspectral Imaging Microscopy
An enhanced dark-field transmission optical microscope (Olympus BX41)
equipped with a hyperspectral imaging (HSI) spectrometer (CytoViva® Hyperspectral
Imaging System, Auburn, AL) was employed to simultaneously visualize fullerenes and
viruses. This was equipped with novel optics that achieve Koehler and Critical
illumination concurrently, which brightens particles 150 times more than conventional
dark field microscopy. Koehler illumination is pre-aligned and fixed whereas Critical
illumination is adjustable (up or down), which aids in finding the focal point including
the position and size of the illuminated spot on the sample. This facilitated observation
of non-diffraction limited optical effects and an improved point spread function
extending the resolution well beyond diffraction limits of a conventional microscope to
obtain spatial details of the sample [266]. The complete composite spectrum of each pixel
(25x25 nm at 100x magnification) in the visible and near infrared wavelengths (400-1000
nm) was collected at 1.5 nm resolution in a hyperspectral image. The spatial distribution
and spectral information of nanoparticles and viruses was derived from each
hyperspectral image using the Environment for Visualization ENVI v.4.4 software.
Samples were stained with 2.5% phosphotungstic acid (10 µL to 1 mL sample) for 30
70
seconds and 10 µL aliquots were imaged using glass slides. All images were acquired at
a constant high gain and a 250 ms exposure time and background subtracted using ten
dark-current images obtained prior to the sample scan. Using n-dimensional visualizer,
the most spectrally active materials (i.e. endmembers) were identified, then compared
with the spectral library of the materials, and finally using mixture tuned matched
filtering technique, individual endmembers were mapped in the sample’s hyperspectral
image. The endmembers (spectral library plots) for pure and mixtures of MS2 and
fullerenes were derived from hyperspectral images using ENVI’s Spectral Hourglass
algorithm (see Appendix A for HSI details, Figure 40). Three to five images with a field
of view of 350-400 mm for each sample was examined and analyzed.
4.1.3 Results and Discussion
4.1.3.1 1O2 Generation by UV-A Sensitized Fullerenes
1O2 production was determined by comparing the measured fluorescence of
SOSG in suspensions versus a Rose Bengal standard curve. Standard suspensions of
Rose Bengal, a known 1O2 sensitizer (Φ = 0.75 [267]), were exposed to UV-A light
centered at 365 nm. From the irradiance and reaction vial geometry, the cumulative 1O2
generated was determined from the FSU measurements (see Appendix A for more
details).
71
Page%6%
ENVI% v.4.4% software.% Samples% were% stained%with% 2.5%% phosphotungstic% acid% (10% μL% to% 1%mL% sample)% for% 30%127"
seconds%and%10%μL%aliquots%were% imaged%using%glass%slides.%All% images%were%acquired%at%a%constant%high%gain%128"
and%a%250%ms%exposure%time%and%background%subtracted%using%ten%dark\current%images%obtained%prior%to%the%129"
sample% scan.% Using% n\dimensional% visualizer,% the% most% spectrally% active% materials% (i.e.% endmembers)% were%130"
identified,%then%compared%with%the%spectral% library%of%the%materials,%and%finally%using%mixture%tuned%matched%131"
filtering% technique,% individual% endmembers% were% mapped% in% the% sample’s% hyperspectral% image.% The%132"
endmembers% (spectral% library% plots)% for% pure% and% mixtures% of% MS2% and% fullerenes% were% derived% from%133"
hyperspectral%images%using%ENVI’s%Spectral%Hourglass%algorithm%(see%SI%for%HSI%details%Figure%S6).%Three%to%five%134"
images%with%a%field%of%view%of%350\400%µm%for%each%sample%was%examined%and%analyzed.%135"
RESULTS,AND,DISCUSSION,136"1O2, Generation, by,UV4A, Sensitized, Fullerenes.! 1O2%production*was*determined*by* comparing* the*measured*137"
fluorescence* of* SOSG* in* suspensions* versus* a* Rose* Bengal* standard* curve.* Standard* suspensions* of* Rose%138"
Bengal," a" known" 1O2%sensitizer( (Φ! =" 0.75" [25]),# were# exposed# to#UVA% light% centered% at% 365% nm.% From% the%139"
irradiance% and% reaction% vial% geometry,% the% cumulative% 1O2% generated% was% determined% from% the% FSU%140"
measurements%(see%SI%for%more%details).%%141"
,142"Figure,1.%1O2%generation%rates%in%UV\A%sensitized%fullerene%nanoparticle%suspensions%(15%mM%PBS).%143"
As%summarized% in%Figure%1,%UV\A\sensitized%C60(NH2)6%generated%1O2%at% the%highest%rate% (0.969±0.039%144"
µM% min\1)% followed% by% C60(OH)24% (0.686±0.011% µM% min\1),% C60(OH)6% (0.563±0.014% µM% min\1),% and% aqu\nC60%145"
0 1 2 3 4 50
1
2
3
4
5
aqu-nC60: 0.132±0.004
C 60(OH) 6
: 0.563±0.014C 60
(OH) 24: 0.686±0.011
C 60(N
H 2) 6: 0
.969±
0.039
Cum
ulat
ive
1 O2 c
once
ntra
tion
(µM
)
Exposure time (minute)
C60(OH)24
C60(OH)6
aqu-nC60
C60(NH2)6
Figure 13: 1O2 generation rates in UV-A sensitized FNP suspensions (15 mM PBS)
As summarized in Figure 13, UV-A-sensitized C60(NH2)6 generated 1O2 at the
highest rate (0.969±0.039 µM min-1) followed by C60(OH)24 (0.686±0.011 µM min-1),
C60(OH)6 (0.563±0.014 µM min-1), and aqu-nC60 (0.132±0.004 µM min-1). 1O2 production
depends on the quantum yield of primary nanoparticles and aggregate morphology
[219, 268]. Shielding by particles on the outer surface of aggregates reduces absorption of
the excitation radiation by those in the interior (an effect which becomes more
pronounced for denser aggregates). On a per monomer basis, this shielding leads to a
decreased rate of optically induced transitions and ultimately, lower ROS production.
Additionally, for light absorbed by particles in the interior of the aggregate, a more
dense morphology will increase the probability of physical and chemical quenching,
which also lowers ROS production [254].
72
Page%7%
(0.132±0.004%µM%min\1).%1O2%production%depends%on%the%quantum%yield%of%primary%nanoparticles%and%aggregate%146"
morphology% [26,% 27].% Shielding% by% particles% on% the% outer% surface% of% aggregates% reduces% absorption% of% the%147"
excitation% radiation% by% those% in% the% interior% (an% effect% which% becomes% more% pronounced% for% denser%148"
aggregates).%On%a%per%monomer%basis,%this%shielding%leads%to%a%decreased%rate%of%optically%induced%transitions%149"
and% ultimately,% lower% ROS% production.% Additionally,% for% light% absorbed% by% particles% in% the% interior% of% the%150"
aggregate,%a%more%dense%morphology%will% increase%the%probability%of%physical%and%chemical%quenching,%which%151"
also%lowers%ROS%production%[10].%%152"
,153"Figure,2.%TEM%images%depicting%the%morphology%of%nanoparticles.%aqu\nC60%crystal%lattice%planes%are%visible%in%the%inset%of%154"(a)%labeled%(e).%For%Figure%e,%scale:%7%mm=20%nm.,155"
For%C60(OH)6% (number%weighted%average%size%of%8.46±0.22%nm),%we%see%an% increase% in% 1O2%generation%156"
compared%with% aqu\nC60% (average% size% of% 80.50±0.46% nm)% despite% a%more% disturbed% π\electron% shell% arising%157"
from%functionalization.%This%is%likely%because%hydroxylation%results%in%a%smaller,%more%open%aggregate%compared%158"
with% the% crystalline% structure% that% is% seen% for% nC60% in% TEM% images% (Figure% 2a% and% e).% The% fact% that% the% 1O2%159"
generation%has%not%overtaken%that%of%the%highly%dendritic%fullerol%aggregates%further%illustrates%the%importance%160"
of%extent%of% functionalization%and%aggregate%morphology.%The%addition%of% six%amine% functional%groups% to% the%161"
fullerene%cage%appears%to%create%a%more%loosely%associated%configuration%similar%to%C60(OH)24,%as%can%be%seen%in%162"
Figures%2%(c)%and%(d),%while%doing%so%with%a%lesser%degree%of%π\bond%interruption,%which%further%illustrates%the%163"
Figure 14: TEM images depicting the morphology of nanoparticle aggregates. Aqu-nC60 crystal lattice planes are visible in (e).
For C60(OH)6 (number weighted average size of 8.46±0.22 nm), we see an increase
in 1O2 generation compared with aqu-nC60 (average size of 80.50±0.46 nm) despite a
more disturbed π-electron shell arising from functionalization. This is likely because
hydroxylation results in a smaller, more open aggregate compared with the crystalline
structure that is seen for nC60 in TEM images (Figure 14 (a) and (e)). The fact that the 1O2
generation has not overtaken that of the highly dendritic fullerol aggregates further
illustrates the importance of extent of functionalization and aggregate morphology. The
addition of six amine functional groups to the fullerene cage appears to create a more
loosely associated configuration similar to C60(OH)24, as can be seen in Figure 14 (c) and
(d), while doing so with a lesser degree of π-bond interruption, which further illustrates
73
the importance of the type of functional group. This combination, in addition to a
smaller aggregate size (average size of 1.92±0.46 nm) compared to C60(OH)24 (average
size of 4.46±0.38 nm), results in the highest 1O2 production rate of all nanoparticles
investigated.
4.1.3.2 1O2 Mediated MS2 Inactivation
In suspensions of MS2 + fullerenes + 90 µM β-carotene (an efficient 1O2
quencher), the inactivation rates were similar to the rate measured with UV-A alone (see
Appendix A, Table 7). Also, the production of superoxide, as measured by XTT (2,3-
bis(2-methoxy-4- nitro-5-sulfophenyl)-2H-tetrazolium-5- carboxanilide) was negligible
in all suspensions. Hence, exogenous 1O2 was predominantly responsible for enhanced
MS2 inactivation rates in UV-A sensitized fullerene suspensions. This result was
expected given the lack of electron donor needed for type I sensitization and confirms
previous results [246].
Virus inactivation in illuminated nanoparticle suspensions closely followed first
order kinetics (Figure 15). aqu-nC60, C60(OH)6, C60(OH)24, and C60(NH2)6 increased MS2
inactivation rates by a factor of 1.7, 5.1, 5.5, and 11.4 respectively compared with UV-A
alone following the same trend as 1O2 production. No dark inactivation demonstrated
that the FNPs investigated herein were not directly virucidal (Appendix A, Table 8).
Inactivation by UV-A alone could be due to the formation of cross-linkages via disulfide
bonds in Cys and Trp photooxidation, (Cys and Trp each constitute ~ 1.6% of the capsid)
74
Page%8%
importance%of%the%type%of%functional%group.%This%combination,%in%addition%to%a%smaller%aggregate%size%(average%164"
size% of% 1.92±0.46% nm)% compared% to% C60(OH)24% (average% size% of% 4.46±0.38% nm),% results% in% the% highest% 1O2%165"
production%rate%of%all%nanoparticles%investigated.%166"1O2, Mediated, MS2, Inactivation.% In% suspensions% of% MS2% +% fullerenes% +% 90% µM% β\carotene% (an% efficient% 1O2%167"
quencher),%the%inactivation%rates%were%similar%to%the%rate%measured%with%UV\A%alone%(see%SI%Table%S2).%Also,%the%168"
production% of% superoxide,% as% measured% by% XTT% (2,3\bis(2\methoxy\4\% nitro\5\sulfophenyl)\2H\tetrazolium\5\%169"
carboxanilide)% was% negligible% in% all% suspensions.% Hence,% exogenous% 1O2% was% predominantly% responsible% for%170"
enhanced%MS2%inactivation%rates%in%UV\A%sensitized%fullerene%suspensions.%This%result%was%expected%given%the%171"
lack%of%electron%donor%needed%for%type%I%sensitization%and%confirms%previous%results%[1].%172"
,173"Figure,3.%MS2%inactivation%by%UV\A%alone%and%UV\A%sensitized%aqu\nC60,%C60(OH)6,%C60(OH)24,%and%C60(NH2)6.%174"
Virus% inactivation% in% illuminated%nanoparticle%suspensions%closely% followed% first%order%kinetics% (Figure%175"
3).%aqu\nC60,%C60(OH)6,%C60(OH)24,%and%C60(NH2)6%increased%MS2%inactivation%rates%by%a%factor%of%1.7,%5.1,%5.5,%and%176"
11.4%respectively%compared%with%UV\A%alone%following%the%same%trend%as%1O2%production.%No%dark%inactivation%177"
demonstrated%that%the%fullerenes%investigated%herein%were%not%directly%virucidal%(Table%S3).%Inactivation%by%UV\178"
A%alone%could%be%due%to%the%formation%of%cross\linkages%via%disulfide%bonds%in%Cys%and%Trp%photooxidation,%(Cys%179"
and%Trp%each%constitute%~%1.6%%of%the%capsid)%[28,%29].%The%photoinactivation%rate%with%6.9%µM%C60(OH)24%in%this%180"
study%(0.187%min\1)%was%significantly%higher%than%that%obtained%in%our%earlier%using%40%µM%C60(OH)24%(0.102%min\181"1)% indicating% that% looser% and% smaller% fullerol% aggregates% generated% higher% 1O2% concentrations% and/or% more%182"
0 1 2 3 4 5-2.5
-2.0
-1.5
-1.0
-0.5
0.0 y =-(0.034±0.001)xy =-(0.059±0.003)x
y =-(0.187±0.006)x
UVA aloneaqu-nC60+UVAC60(OH)6+UVAC60(NH2)6+UVAC60(OH)24+UVA Dark
y =-(0.389±0.012)x
y =-(0.175±0.007)x
Log
(Nt/N
0)
Exposure time (minute)
Figure 15: MS2 inactivation by UV-A alone and UV-A sensitized aqu-nC60, C60(OH)6, C60(OH)24, C60(NH2)6.
[269, 270]. The photoinactivation rate with 6.9 µM C60(OH)24 in this study (0.187 min-1)
was significantly higher than that obtained in our earlier using 40 mM C60(OH)24 (0.102
min-1) indicating that looser and smaller fullerol aggregates generated higher 1O2
concentrations and/or more closely associated with MS2. (Note that in this study, a 100
nm filtration step was introduced to remove very large aggregates.) Biochemical
transformations in viruses were probed next.
4.1.3.3 Alterations to Protein Secondary Structures
Peak assignments in the mid-infrared region (1800-1000 cm-1) are summarized in
Appendix A, Table 9. 1O2-induced changes to protein secondary structures in MS2 after 5
75
minutes of exposure were quantified from relative peak areas in curve fitted amide I
spectra in the region 1700-1600 cm-1 (Figure 16).
Page%9%
closely%associated%with%MS2.%(Note%that%in%this%study,%a%100%nm%filtration%step%was%introduced%to%remove%very%183"
large%aggregates.)%Biochemical%transformations%in%viruses%were%probed%next.%184"
Alterations, to, Protein, Secondary, Structures.,Peak%assignments% in% the%mid\infrared% region% (1800\1000% cm\1)%185"
are%summarized%in%SI%Table%S4.%1O2\induced%changes%to%protein%secondary%structures%in%MS2%after%5%minutes%of%186"
exposure%were%quantified%from%relative%peak%areas%in%curve%fitted%amide%I%spectra%in%the%region%1700\1600%cm\1%187"
(Figure%4).%%188"
,189"Figure,4.%Changes%in%protein%secondary%structures%(intermolecular%extended%chains,%β\sheets%and%random%coils,%α\%and%3\190"turn%helices,%and%antiparallel%β\sheets%and%aggregated%strands%following%5\minute%exposure%to%UV\A%illuminated%fullerene%191"suspensions%from%ATR\FTIR.%CT%values%correspond%to%10\9%mg\min/L.%192"
Photoilluminated% nanoparticles% increased% peak% areas% of% α\% and% 3\turn% helices% and% antiparallel%193"
aggregated% strands/β\sheets% but% decreased% infrared% vibrational% freedom% of% intermolecular% extended% chains%194"
and% β\sheets/random% coils% in% viruses.% Changes% in% the% interactions% between% α\helical% structures% and% β\195"
structures%suggests%unfolding%of%subunits%of%coat%proteins%and%capsid%destabilization%via%deformation%or%rupture%196"
[30].%Bonds%in%intermolecular%extended%chains,%β\sheets,%and%random%coils%were%potentially%strained%resulting%197"
in% the% loss%of%molecular%tensional%energy%needed%for%maintaining%structural% integrity%of% the%capsid% [31].%Peak%198"
shifts%associated%with% intermolecular%chains%also% indicate%morphological%disorder% in% the%capsid% (SI%Figure%S7).%199"
Since%the%A\protein%assists%in%storing%tension%or%strain%energy%in%the%coat%protein%[31],%MS2%inactivation%might%200"
also%be%associated%with%reduced%ability%to%inject%E.,coli%with%its%genomic%RNA.%201"
0.0
0.2
0.4
0.6
0.8
1.0
1.2
CT=7.89 CT=10.8CT=6.22CT=1.61
CT=0 Intermolecular extended chains β-sheets/random coils α- & 3-turn helices Antiparallel aggregated strands/β-sheets
MS2 +C60(NH2)6
MS2 +C60(OH)24
MS2 +C60(OH)6
MS2 +aqu-nC60
MS2 alone
Ar
ea u
nits
(-)
Figure 16: Changes in protein secondary structures (intermolecular extended chains, β-sheets and random coils, α- and 3-turn helices, and antiparallel β-sheets and
aggregated strands following 5 minute exposure to UV-A illuminated FNP suspensions from ATR-FTIR. CT values correspond to 10-9 mg min L-1.
Photoilluminated nanoparticles increased peak areas of α- and 3-turn helices and
antiparallel aggregated strands/β-sheets but decreased infrared vibrational freedom of
intermolecular extended chains and β-sheets/random coils in viruses. Changes in the
interactions between α-helical structures and β-structures suggests unfolding of
subunits of coat proteins and capsid destabilization via deformation or rupture [271].
Bonds in intermolecular extended chains, β-sheets, and random coils were potentially
strained resulting in the loss of molecular tensional energy needed for maintaining
structural integrity of the capsid [272]. Peak shifts associated with intermolecular chains
also indicate morphological disorder in the capsid (Appendix A, Figure 41). Since the A-
76
protein assists in storing tension or strain energy in the coat protein [272], MS2
inactivation might also be associated with reduced ability to inject E. coli with its
genomic RNA.
4.1.3.4 Formation of Carbonyl Groups and Protein Conformational Changes
Carbonyl groups, which appear between 1740 and 1730 cm-1 of the mid-infrared
spectrum were detected in all viruses exposed to photoactivated fullerenes (but not UV-
alone) demonstrating oxidation of capsid proteins as seen in Figure 17 [273, 274].
Page%10%
Formation,of,Carbonyl,Groups,and,Protein"Conformational,Changes.%Carbonyl%groups,%which%appear%between%202"
1740% and% 1730% cm\1% of% the% mid\infrared% spectrum% were% detected% in% all% viruses% exposed% to% photoactivated%203"
fullerenes%(but%not%UV\alone)%demonstrating%oxidation%of%capsid%proteins%as%seen%in%Figure%5%[32,%33].%%204"
,,
Figure, 5.% (a)% Evolution% of% carbonyl% groups% in%MS2% upon% 1O2% oxidation% from% ATR\FTIR.% (b)% The% extent% of%MS2% proteins%205"oxidation%by%1O2%at%the%end%of%5%min%of%UV\A%sensitization%is%shown%here%in%terms%of%carbonyl%concentrations%from%OxyBlot%206"assay.%CT%values%correspond%to%10\9%mg\min/L.%207"
Total% carbonyl% groups% relative% peak% areas% (aqu\nC60% 0.0088,% C60(OH)6% 0.0316,% C60(OH)24% 0.0337% and%208"
C60(NH2)6%0.0523%area%units)%evolved%monotonically%with%CT%(Figure%5a,%details%of%CT%calculations%are%in%the%final%209"
section%of%the%manuscript%and%SI).%Importantly,%since%no%carbonyl%groups%were%detected%in%viruses%exposed%to%210"
fullerenes% in% the% dark,% UV\A% alone,% or% photoactivated% fullerenes% +% β\carotene,% protein% oxidation% is% directly%211"
related% to% exogenous% 1O2% delivered% to% the% capsid% and% the% degree% of% inactivation% (Figure% 5b).%SDS\PAGE% and%212"
Western%blot%analysis%as%described%by%the%OxyBlot%assay%allowed%the%quantification%of%carbonyl%content%of%A\%213"
and% coat\protein.% As% summarized% in% Figure% 5b,% it% appears% that% carbonyl% formation% in% A\protein% stopped%214"
increasing%significantly%beyond%a%CT%of%only%1.61x10\9%mg%min/L%(C60(OH)6)%whereas%it%continued%to%increase%in%215"
coat% proteins% consistent%with% its%much% higher% protein% content% and% consequent% greater% availability% of% amino%216"
acid%residues%for%reactions%with%1O2.%Note%that%each%MS2%virion%has%180%copies%of%the%coat%protein%but%only%one%217"
copy%of%the%A\protein.%The%highest%degree%of%oxidation%occurred%in%C60(NH2)6%suspensions%(A\protein:%1.03±0.11%218"
nanomoles% carbonyls/mg% protein% and% coat% protein:% 4.23±0.60% nanomoles% carbonyls/mg% protein).% Carbonyl%219"
group%content%in%both%A\protein%and%coat%protein%increased%with%1O2%doses%confirming%FTIR%results%(Figure%5b).%220"
7%%of%the%amino%acids%in%the%MS2%capsid%are%highly%reactive%with%1O2%(His,%Trp,%Tyr,%Met,%and%Cys;%SI%Table%S5),%221"
1760 1750 1740 1730 1720
MS2+C60(OH)24
Increasingcarbonyl content
MS2alone
MS2+aqu-nC60
MS2+C60(NH2)6
MS2+C60(OH)6
Abs
orba
nce
units
(-)
Wavenumber (cm-1)
0
1
2
3
4
5 CT=10.8
CT=7.89
CT=6.22
CT=1.61
CT=0CT=0CT=0
C60(NH2)6+β-carotene
UV-A C60(NH2)6C60(OH)24
C60(OH)6aqu-nC60Dark +
fullerenes
Car
bony
l con
tent
(nm
ol/m
g pr
otei
n)
A-protein Coat protein
a,
b,
Figure 17: (a) Evolution of carbonyl groups in MS2 upon 1O2 oxidation from ATR-FTIR. (b) The extent of MS2 protein oxidation by 1O2 at the end of 5 min of UV-A sensitization is shown here in terms of carbonyl concentrations from the OxyBlot
assay. CT values correspond to 10-9 mg min L-1.
Total carbonyl groups relative peak areas (aqu-nC60 0.0088, C60(OH)6 0.0316,
C60(OH)24 0.0337 and C60(NH2)6 0.0523 area units) evolved monotonically with CT (Figure
17 (a), details of CT calculations are in the final section of this Chapter and Appendix A).
Importantly, since no carbonyl groups were detected in viruses exposed to fullerenes in
the dark, UV-A alone, or photoactivated fullerenes + β-carotene, protein oxidation is
directly related to exogenous 1O2 delivered to the capsid and the degree of inactivation
77
(Figure 17 (b)). SDS-PAGE and Western blot analysis as described by the OxyBlot assay
allowed the quantification of carbonyl content of A- and coat-protein. As summarized in
Figure 17 (b), it appears that carbonyl formation in A-protein stopped increasing
significantly beyond a CT of only 1.61x10-9 mg min L-1 (C60(OH)6) whereas it continued to
increase in coat proteins consistent with its much higher protein content and consequent
greater availability of amino acid residues for reactions with 1O2. Note that each MS2
virion has 180 copies of the coat protein but only one copy of the A-protein. The highest
degree of oxidation occurred in C60(NH2)6 suspensions (A-protein: 1.03±0.11 nanomoles
carbonyls mg-1 protein and coat protein: 4.23±0.60 nanomoles carbonyls mg-1 protein).
Carbonyl group content in both A-protein and coat protein increased with 1O2 doses
confirming FTIR results (Figure 17 (b)). 7% of the amino acids in the MS2 capsid are
highly reactive with 1O2 (His, Trp, Tyr, Met, and Cys; Appendix A, Table 10), which are
also present in the A-protein that is crucial for infection. SDS PAGE confirmed
conformational changes such as unfolding of subunits or cross-linkages (slower
migration of bands in lanes D, E, F, G and I in Appendix A, Figure 42) in A-protein and
coat proteins of MS2 across all treatments including UV-A alone. Oxidative damages to
the A-protein suggest alterations in MS2-host interactions attributing inactivation to the
loss of infectivity similar to DNA viruses such as PRD1 and T7 [247].
78
4.1.3.5 Evidence for Capsid Rupture or Deformation
As seen in Figure 18, certain amino acids (Pro, His, Trp, and Ser) peaks and RNA
appeared in spectra of MS2 exposed to illuminated FNPs but not UV-A alone. This
suggests capsid rupture/deformation and the broadened peaks in the region 1190-1130
cm-1 likely indicate oxidation of exposed RNA. The appearance of new peaks show
localized capsid damages potentially at the locations of Pro, His, Trp, and Ser on the
capsid, or even rupture as seen in Figure 19 (e).
Page%11%
which% are% also% present% in% the% A\protein% that% is% crucial% for% infection.% SDS% PAGE% confirmed% conformational%222"
changes%such%as%unfolding%of%subunits%or%cross\linkages%(slower%migration%of%bands%in%lanes%D,%E,%F,%G%and%I%in%SI%223"
Figure% S8)% in% A\protein% and% coat% proteins% of% MS2% across% all% treatments% including% UV\A% alone.% Oxidative%224"
damages% to% the%A\protein% suggest% alterations% in%MS2\host% interactions%attributing% inactivation% to% the% loss%of%225"
infectivity%similar%to%DNA%viruses%such%as%PRD1%and%T7%[2]."%226"
Evidence,for,Capsid,Rupture,or,Deformation.%As%seen%in%Figure%6,%certain%amino%acids%(Pro,%His,%Trp,%and%Ser)%227"
peaks% and% RNA% appeared% in% spectra% of% MS2% exposed% to% illuminated% fullerenes% but% not% UV\A% alone.% This%228"
suggests% capsid% rupture/deformation% and% the% broadened% peaks% in% the% region% 1190\1130% cm\1% likely% indicate%229"
oxidation% of% exposed% RNA.% The% appearance% of% new% peaks% show% localized% capsid% damages% potentially% at% the%230"
locations%of%Pro,%His,%Trp,%and%Ser%on%the%capsid,%or%even%rupture%as%seen%in%Figure%7e.%%231"
%232"Figure,6.%ATR\FTIR%spectra%of%MS2%exposed%to%UV\A%illuminated%fullerenes%showing%development%of%amino%acid%residues%233"and%RNA%compared%to%UV\A%alone.%%234"
Direct,Qualitative,Evidence,for,Differential,Association,of,Different,Fullerenes,with,MS2.%Figures%7a\d%show%235"
representative%5\fold%zooms%of%mixture\tuned%matched%filtering%(MTMF)%results%from%HSI%analysis%of%MS2%with%236"
C60(NH2)6,% aqu\nC60% C60(OH)6,% and% C60(OH)24% respectively% in% aqueous% media.% The% following% trends% were%237"
consistently%observed;%C60(NH2)6%and%MS2%were%closest%to%each%other,%the%polyhydroxylated%fullerols%and%MS2%238"
were%most%separated,%whereas%aqu\C60%was%at%intermediate%distances%from%MS2.%These%images%were%obtained%239"
in% the%wet% phase%with% limited% sample%preparation% thereby% closely% capturing% conditions%during%nanoparticles%240"
1500 1400 1300 1200 1100 1000
MS2+C60(OH)
24
SerTr
pTrpHisP
ro
Trp
/ Tyr
-O-
RN
A P
O2-
RN
A C
-H
RN
A ri
bose
RN
A ri
bose
RN
A P
O2-
RN
A ri
bose
RN
A ri
bose
MS2
MS2+aqu/nC60
MS2+C60(NH
2)6
MS2+C60(OH)
6
1120
1043
106011
6011
8412
1812
44
1273
-6
1323
137714
05144914
74
Abs
orba
nce
Uni
ts (A
.U.)
Wavenumber (cm-1)
Figure 18: ATR-FTIR spectra of MS2 exposed to UV-A illuminated FNPs showing the development of amino acid residues and RNA compared to UV-A alone.
4.1.3.6 Direct Qualitative Evidence for Differential Association of FNPs with MS2
Figure 19 (a)-(d) show representative 5-fold zooms of mixture-tuned matched
filtering (MTMF) results from HSI analysis of MS2 with C60(NH2)6, aqu-nC60 C60(OH)6,
and C60(OH)24 respectively in aqueous media. The following trends were consistently
79
observed; C60(NH2)6 and MS2 were closest to each other, the polyhydroxylated fullerols
and MS2 were most separated, whereas aqu-C60 was at intermediate distances from MS2.
These images were obtained in the wet phase with limited sample preparation thereby
closely capturing conditions during nanoparticles photoactivation and subsequent virus
inactivation. Although not conclusive because of the necessary sample dehydration,
TEM imagery (Figure 19 (e)-(h)) confirms HSI results where C60(NH2)6 nanoparticles
were present closer to MS2 compared with aqu-nC60, C60(OH)6, and C60(OH)24.
Page%12%
photoactivation%and%subsequent%virus% inactivation.%Although%not%conclusive%because%of%the%necessary%sample%241"
dehydration,% TEM% imagery% (Figures% 7e\h)% confirms% HSI% results% where% C60(NH2)6% nanoparticles% were% present%242"
closer%to%MS2%compared%with%aqu\nC60,%C60(OH)6,%and%C60(OH)24.%243"
% % % %
% % % %
Figure,7.%Relative%proximity%of%various%fullerenes%with%MS2%following%MTMF%analysis%of%hyperspectral%images.%5\fold%zoom\244"ins% (1%cm%=%20%µm)%of%noise\free% image%data%with%superimposed%classification%results%are%shown.%MS2%viruses%appear%as%245"green% pixels% after% staining% with% 2.5%% phosphotungstic% acid.% C60(NH2)6% is% shown% as% red% color% (a),% aqu\nC60% as% blue% (b),%246"C60(OH)6%as%purple%(c),%and%C60(OH)24%as%pink%(d).%Corresponding%TEM%images%after%sample%dehydration%are%shown%in%(e\h).%247"
Under% our% experimental% conditions,% the% electrophoretic% mobilities% of% MS2,% C60(NH2)6,% aqu\nC60,%248"
C60(OH)6,%and%C60(OH)24%were%\1.09±0.09,% \1.09±0.10,% \1.76±0.29,% \3.65±0.31,%and%\3.57±0.30%(x10\8%m2%V\1%s\1),%249"
respectively.%ζ%potentials%were%then%calculated%from%EPM%using%the%Henry%equation%(SI%Table%S1).%The%observed%250"
trends% in% separation% distances% are% therefore% consistent% with% electrostatic% interactions% with% proximity% of%251"
nanoparticles% to%phages% increasing%with%decreasing%magnitude%of% the% ζ% potential.%Greater% affinity% allows% the%252"
more% efficient% transfer% of% 1O2% from% nanoparticles% to% viruses,% especially% given% its% short% lifetime% thereby%253"
facilitating% inactivation.% Capsid% rupture% and% deformations% seen% in% Figures% 7e\h% are% attributed% to% oxidative%254"
damages%by%1O2.%,255"1O2,Delivery,and,Proximity, Implications, for,Chick4Watson,Kinetics:,Log%virus%survival%ratios%were%quantified%256"
using% the% Chick\Watson% model% reported% for% conventional% disinfectants% [34]% and% cationic% fullerenes% [16].%257"
Measurements%of%fluorescence!over%time%represent%a%cumulative%production%of%1O2."The"change"in"fluorescence"258"
b, c,a,
e, h,f, g,
d,
Figure 19: Relative proximity of various fullerenes with MS2 following MTMF analysis of hyperspectral images. 5-fold zoom-ins (1 cm = 20 µm) of noise-free image
data with superimposed classification results are shown. MS2 viruses appear as green pixels after staining with 2.5% phosphotungstic acid. (a) C60(NH2)6 is shown in red, (b)
aqu-nC60 in blue, (c) C60(OH)6 in purple, and (d) C60(OH)24 in pink. (e) –(h) Corresponding TEM images after sample dehydration.
Under our experimental conditions, the electrophoretic mobilities of MS2,
C60(NH2)6, aqu-nC60, C60(OH)6, and C60(OH)24 were -1.09±0.09, -1.09±0.10, -1.76±0.29, -
3.65±0.31, and -3.57±0.30 (x10-8 m2 V-1 s-1), respectively. ζ potentials were then calculated
80
from EPM using the Henry equation (Appendix A, Table 6). The observed trends in
separation distances are therefore consistent with electrostatic interactions with
proximity of nanoparticles to phages increasing with decreasing magnitude of the ζ
potential. Greater affinity allows the more efficient transfer of 1O2 from nanoparticles to
viruses, especially given its short lifetime thereby facilitating inactivation. Capsid
rupture and deformations seen in Figure 19 (e)-(h) are attributed to oxidative damage by
1O2.
4.1.3.7 1O2 Delivery and Proximity Implications for Chick-Watson Kinetics
Log virus survival ratios were quantified using the Chick-Watson model
reported for conventional disinfectants [275] and cationic fullerenes [33]. Measurements
of fluorescence over time represent a cumulative production of 1O2. The change in
fluorescence over time therefore represents the instantaneous 1O2 consumption
following the second order reaction between SOSG and 1O2, presumed to take place near
the FNP surface where the 1O2 concentration is [1O2]0 such that:
dFSUdt
= k1O2 [1O2 ]0[SOSG] (Eq 25.)
where FSU is the detected fluorescence resulting from the cumulative reaction of SOSG
with the produced 1O2 over the irradiance time, T, and 2
1Ok is a second order rate
constant. Using a sufficiently large SOSG concentration compared with 1O2 and short
irradiance times gives the pseudo first order equation:
81
dFSUdt
= k '1O2 [1O2 ]0
(Eq 26.)
where 2
1' Ok = 5.8x109 FSU mM-1 min-1, determined experimentally for our conditions (see
Appendix A). Integrating, an expression for a pseudo-CT is obtained in terms of the time
required to achieve a given level of inactivation and the steady state 1O2 concentration
probed by SOSG on or very near the surface of fullerene:
CT = FSU(t)k '1O2
= [1O2 ]0 t (Eq 27.)
MS2 log survival ratios corresponding to the four nanoparticles evaluated in this
study all conformed to Chick-Watson kinetics with all data falling on a single straight
line in Figure 20. This suggests that similar inactivation mechanisms were associated
with each nanoparticle. However, the 1O2 concentration near the virus capsid, and thus
the time required to achieve a given level of inactivation may vary with proximity
between fullerenes and viruses. 1O2 decays rapidly with distance from a photosensitizer
source, decreasing in concentration for example, by over an order of magnitude after
only approximately 10 nm for the case of humic acid [67]. This implies that for a given
measured surface concentration of 1O2 there are many combinations of separation
distance between virus and fullerene and the associated time required to produce a
given level of inactivation that will satisfy the constraint that CT is constant for all
fullerene types. In other words, the effective concentration that the virus is exposed to
should depend not only on the generation rate of 1O2 but also the separation distance
82
between the nanoparticles and the viruses. Inactivation rates followed the order: aqu-
nC60 < C60(OH)6 ≈ C60(OH)24 < C60(NH2)6 (Figure 15), which could be due to either
differences in 1O2 production rates arising from changes in the quantum yield and
aggregate structure or differences in proximity between the fullerenes and MS2, both of
which followed similar trends. EPM measurements and visual evidence by HSI and
TEM indicate that the derivatization of the fullerene cage affected the affinity of the
particle for the viral surface. For example, C60(NH2)6 was not only the most copious 1O2
producer but it also associated most closely with MS2, leading to its highest efficacy for
virus inactivation in Figure 15. Thus, when examining inactivation rates in this study,
the importance of particle morphology could not be differentiated from virus proximity
as all FNPs conformed to the same CT curve seen in Figure 20. Had virus log survival
ratio differed for the same CT dose, it would be possible to distinguish between the two
factors contributing to inactivation, i.e. 1O2 concentration and proximity.
The only other examination of the applicability of Chick-Watson kinetics to MS2
inactivation by nanoparticles [33], used a derivatized cationic fullerene and reported CT
values that are orders of magnitude greater than our measurements. Differences in CT
between the two studies are likely to arise given the effects of proximity and
83
Page%14%
the%order:%aqu\nC60%<%C60(OH)6%≈%C60(OH)24%<%C60(NH2)6%(Figure%3),%which%could%be%due%to%either%differences%in%1O2%282"
production%rates%arising%from%changes%in%the%quantum%yield%and%aggregate%structure%or%differences%in%proximity%283"
between% the% fullerenes% and% MS2,% both% of% which% followed% similar% trends.% EPM% measurements% and% visual%284"
evidence% by% HSI% and% TEM% indicate% that% the% derivatization% of% the% fullerene% cage% affected% the% affinity% of% the%285"
particle% for% the%viral% surface.%For%example,%C60(NH2)6%was%not%only% the%most% copious% 1O2%producer%but% it% also%286"
associated%most%closely%with%MS2,%leading%to%its%highest%efficacy%for%virus%inactivation%in%Figure%3.%Thus,%when%287"
examining% inactivation%rates% in%this%study,%the% importance%of%particle%morphology%could%not%be%differentiated%288"
from%virus%proximity%as%all%fullerenes%conformed%to%the%same%CT%curve%seen%in%Figure%8.%Had%virus%log%survival%289"
ratio%differed%for%the%same%CT%dose,%it%would%be%possible%to%distinguish%between%the%two%factors%contributing%to%290"
inactivation,%i.e.%1O2%concentration%and%proximity.%%291"
,292"Figure,8.%Evaluation%of%the%Chick\Watson%model%to%describe%MS2$inactivation$kinetics(by(all(four(nanoparticles.%293"
The% only% other% examination% of% the% applicability% of% Chick\Watson% kinetics% to% MS2% inactivation% by%294"
nanoparticles%[16],%used%a%derivatized%cationic%fullerene%and%reported%CT%values%that%are%orders%of%magnitude%295"
greater%than%our%measurements.%Differences%in%CT%between%the%two%studies%are%likely%to%arise%given%the%effects%296"
of%proximity%and%morphology%on%dose%described%herein.%The%change%in%electrostatic%interactions%arising%from%a%297"
positively%charged% fullerene%versus%negatively%charged% fullerenes%should% favor% increased% inactivation,% though%298"
the% derivatization% used% by% Cho% et% al.% [16]% caused% substantial% aggregation% that%would% be% expected% to% hinder%299"
delivery%of%generated%1O2,%resulting%in%decreased%inactivation.%Aggregates%in%their%work%were%over%1%µm%in%size%300"
0 2 4 6 8 10 12-2.5
-2.0
-1.5
-1.0
-0.5
0.0
k = -(0.1575±0.0062)x10 9
R 2=0.957
C60(OH)6+UVA C60(OH)24+UVA aqu-nC60+UVA C60(NH2)6+UVA
Log
(Nt/N
0)
CT (x 10-9 mg-min/L)
Figure 20: Evaluation of the Chick-Watson model to describe MS2 inactivation kinetics by all four nanoparticles.
morphology on dose described herein. The change in electrostatic interactions arising
from a positively charged fullerene versus negatively charged fullerenes should favor
increased inactivation, though the derivatization used by Cho et al. [33] caused
substantial aggregation that would be expected to hinder delivery of generated 1O2,
resulting in decreased inactivation. Aggregates in their work were over 1 µm in size as
indicated by dynamic light scattering [276] which could both limit transport of
generated 1O2 away from the interior of the aggregate and decrease the total surface area
available to interact with viruses. The uncertainty of these factors in producing the
greater apparent dose necessary to achieve similar inactivation levels between the two
studies highlights the need for further investigation.
84
Additionally, differences in CT could arise from differences in 1O2 detection and
measurement. Indeed, CT values reported by Cho et al. [33] differ significantly not only
from our current study, but also from other estimates [277] reiterating the difficulties in
accurately determining 1O2 concentrations in the proximity of viruses, which is what
effectively determines the extent and kinetics of inactivation. In fact, quantifying
differences in inactivation induced by cationic and anionic fullerenes using a consistent
and uniform 1O2 detection method is the subject of our ongoing research.
Page%15%
as%indicated%by%dynamic%light%scattering%[36]%which%could%both%limit%transport%of%generated%1O2%away%from%the%301"
interior%of%the%aggregate%and%decrease%the%total%surface%area%available%to%interact%with%viruses.%The%uncertainty%302"
of% these% factors% in% producing% the% greater% apparent% dose% necessary% to% achieve% similar% inactivation% levels%303"
between%the%two%studies%highlights%the%need%for%further%investigation.%%304"
Additionally,%differences%in%CT%could%arise%from%differences%in%1O2%detection%and%measurement.%Indeed,%305"
CT%values%reported%by%Cho%et%al.%[16]%differ%significantly%not%only%from%our%current%study,%but%also%from%other%306"
estimates% [37]% reiterating% the% difficulties% in% accurately% determining% 1O2% concentrations% in% the% proximity% of%307"
viruses,% which% is% what% effectively% determines% the% extent% and% kinetics% of% inactivation.% In% fact,% quantifying%308"
differences% in% inactivation% induced% by% cationic% and% anionic% fullerenes% using% a% consistent% and% uniform% 1O2%309"
detection%method%is%the%subject%of%our%ongoing%research.%%%310"
Quantum
'Yield
Derivatization
Fullerene
'Trip
let'Lifetim
e
Affinity'for'virus'surfaces'can'either'increase'or'decrease
1O2 generation
? ?
,311"Figure, 9.% Conceptual% illustration% of% the% interplay% between% fullerene% derivatization,% quantum% yield,% and% 1O2% generation.%312"Though% the%quantitative% shape%of% the% curves%are%not% yet% known,% increasing%derivatization%disrupts% the%pi\electron% shell%313"decreasing%the%quantum%yield%and%1O2%generation.%However,%derivatization%also%leads%to%smaller,%more%open%aggregates,%314"decreasing% the% likelihood% of% triplet\triplet% annihilation% and% self% quenching% and% thereby% increasing% the% lifetime% of% the%315"excited%triplet%state,%resulting%in%greater%1O2%production.%The%observed%
1O2%production%will%thus%be%limited%by%the%quantum%316"yield%for%highly%derivatized%fullerenes%and%by%the%triplet%lifetime%for%lesser%derivatized%fullerenes%such%that%a%maximum%will%317"occur%at%some%optimal%extent%and%type%of%derivatization.%Furthermore,%depending%on%the%derivatization,%the%affinity%for%a%318"virus%surface%may%either%increase%or%decrease.%An%optimal%derivatization%maximizing%1O2%generation%as%well%as%increasing%319"affinity%for%the%biological%target%will%facilitate%inactivation.%%320"
Differences%in%MS2%damage,%separation%distance,%and%1O2%generation%highlight%the%importance%of%both%321"
affinity% and%particle%morphology% (size,% structure,% extent%and% type%of%derivatization)% to% facilitate% the%effective%322"
delivery%of%1O2%to%the%capsid.%Fullerene%cage%derivatization%can%lead%to%a%greater%affinity%for%the%aqueous%phase%323"
Figure 21: Conceptual interplay between fullerene derivatization, quantum yield, and 1O2 generation.
4.1.4 Conclusions
Differences in MS2 damage, separation distance, and 1O2 generation highlight the
importance of both affinity and particle morphology (size, structure, extent and type of
derivatization) to facilitate the effective delivery of 1O2 to the capsid. Fullerene cage
derivatization can lead to a greater affinity for the aqueous phase (less hydrophobic),
85
smaller mean aggregate size, and a more open aggregate structure [22, 219, 247]. These
factors tend to favor a longer lifetime for the triplet state of the excited fullerene and
higher rates of 1O2 generation. However, fullerene derivatization also decreases the 1O2
quantum yield [254, 278] and may either increase or decrease its affinity to the viral
surface.
These considerations suggest that both the extent and the type of derivatization
should be optimized to balance 1O2 generation and fullerene-virus association to achieve
maximum disinfection as illustrated schematically in Figure 21. Though the quantitative
shape of the curves are not yet known, increasing derivatization disrupts the pi-electron
shell, decreasing the quantum yield and 1O2 generation. However, derivatization also
leads to smaller, more open aggregates, reducing the likelihood of triplet-triplet
annihilation and self quenching and thereby increasing the lifetime of the excited triplet
state, resulting in greater 1O2 production. The observed 1O2 production will thus be
limited by the quantum yield for highly derivatized FNPs and by the triplet lifetime for
lesser derivatized FNPs such that a maximum will occur at some optimal extent and
type of derivatization. Furthermore, depending on the derivatization, the affinity for a
virus surface may either increase or decrease. An optimal derivatization maximizing 1O2
generation as well as increasing affinity for the biological target will facilitate
inactivation.
86
4.2 Influence of Inorganic Anions on the Reactivity of TiO2 Nanoparticles
ABSTRACT
The influence of inorganic anions on the photoreactivity and aggregation of
titanium dioxide nanoparticles was assessed by dosing carbonate, chloride, nitrate,
phosphate and sulfate as potassium salts at multiple concentrations. Nanoparticle
stability was monitored in terms of aggregate morphology and electrophoretic mobility
(EPM). Aggregates size and fractal dimension were measured over time by laser
diffraction, and the isoelectric point (IEP) as a function of anion and concentration was
obtained by measuring EPM vs pH. Phosphate, carbonate, and to a lesser extent, sulfate,
all decreased the IEP of TiO2 and stabilized NP suspensions due to specific surface
interactions, whereas this was not observed for nitrate and chloride. TiO2 NPs were
exposed to UV-A radiation and photoreactivity was assessed by monitoring the
production of reactive species over time both at the NP surface (photogenerated holes)
and in the bulk solution (hydroxyl radicals) by observing their reactions with the
selective probe compounds iodide and terephthalic acid, respectively. The generation of
photogenerated holes and hydroxyl radicals was impacted by each inorganic anion to
varying degrees. Carbonate and phosphate inhibited the oxidation of iodide, and this
interaction was successfully described by a Langmuir-Hinshelwood mechanism and
related to the characteristics of TiO2 aggregates. Chloride and nitrate do not specifically
interact with TiO2, and sulfate creates relatively weak interactions with the TiO2 surface
87
such that no decrease in photogenerated hole reactivity was observed. A decrease in
hydroxyl radical generation was observed for all inorganic anions. Quenching rate
constants for the reaction of hydroxyl radicals with each inorganic anion do not provide
a comprehensive explanation for the magnitude of this decrease, which arises from the
interplay of several physicochemical phenomena. This work shows that the reactivity of
NPs will be strongly influenced by the waters they are released to. The impact of anion
species on hydroxyl radical inhibition was as follows: carbonate > chloride > phosphate >
nitrate > sulfate.
4.2.1 Introduction
Titanium dioxide (TiO2) nanoparticles (NPs) are widely used in a multitude of
commercial products, including food, personal care products, paints, coatings, paper,
and fibers due to their brightness, high refractive index, and UV resistance [8, 12, 171,
279]. Applications related to their photoëlectrochemical properties are of particular
research interest involving solar energy conversion, photocatalytic pollution
remediation and photo-induced superhydrophilicity [13, 280]. In particular, TiO2 NP
photocatalysis may be used in water and wastewater treatment, as an advanced
oxidation for chemical pollutants or a disinfection process for pathogens based on the
generation of reactive species when TiO2 NPs in suspension are irradiated by photons of
wavelength below 390 nm [14]. The potential for TiO2 photocatalytic degradation of a
number of contaminants has already been highlighted [281, 282], though further
88
investigations are required to improve its cost-efficiency and to develop its application
at the full scale [283].
The same photochemical properties that make TiO2 appealing from an
engineering standpoint also have implications for NP release into the environment. TiO2
NPs are impacted by the presence of many components of natural waters including
inorganic anions, which are ubiquitous at significant concentrations [29, 284]. The
impact on TiO2 reactivity by inorganic anions can be direct: interference with the
physicochemical processes leading to reactivity, or indirect: changes in the NP stability
and aggregation state which in turn alter reactivity [14]. For the direct mechanism,
several studies have been published indicating that inorganic anions impact the
fundamental physicochemical processes taking place during TiO2 photocatalysis,
including photon absorption, surface adsorption and photocatalytic degradation [285-
292]. Anions compete for the active sites on TiO2 surfaces, which can prevent the
reaction of pollutants that must be adsorbed before degradation. The specific adsorption
of various anions, such as phosphate, leads to a reduction at the TiO2 surface in the
amount of hydroxide ions, which are the precursors of hydroxyl radicals [293-296].
Furthermore, photocatalytic degradation is inhibited when reactive species are directly
quenched by anions. This phenomenon follows the specific reactivity of each compound
with the generated reactive species [297]. Given the difficulty in differentiating the
chemical reactions involved, reactivity assessments have generally been limited to
89
monitoring model contaminant degradation, such as dyes or organic solvents, whose
reaction pathways are complex and mostly unknown.
Similarly, most studies assessing the influence of inorganic ions on the reactivity
of TiO2 suspensions have been carried out without a detailed characterization of TiO2
aggregates that would help to inform the importance of the indirect mechanism. The
influence of inorganic anions on NP aggregation has been extensively documented for
nanoparticle diffusion and transport in aquatic systems [298-302]. This area of research
is of particular interest due to the increasing concern for environmental implications of
nanomaterials. In particular, the interactions among TiO2 nanoparticles can often
effectively be described by DLVO theory, which accounts for van der Waals attraction
and electrical double layer repulsion in determining the potential for aggregation [180,
181, 301]. Generally, the repulsive energy of the electric double layer is dominant for
TiO2 NPs in water. This energy barrier decreases with increasing ionic strength,
corresponding to double layer compression, which decreases electrostatic repulsion. As
a result, the ionic strength of the solution can be related to the stability of a
electrostatically-stabilized NP suspensions [299]. Though some work has looked at the
impact of aggregate size and structure on reactivity, there remains a lack of
understanding of how ionic species and concentration impacts the characteristics of TiO2
suspensions and their reactivity.
90
In this work, the reactivity of TiO2 suspensions in the presence of inorganic
species was assessed at the laboratory scale by monitoring the reaction of reactive
species with probe compounds. Using highly selective chemicals (previously described
in several publications [73, 75, 76]) allows one to assess the specific influence of
inorganic ions on the availability of various reactive species in the development of
radical pathways. Specifically, the oxidation of iodide (dosed as potassium iodide, KI) to
iodine (I2) was used to monitor photogenerated holes (h+), and hydroxylation of
terephthalic acid (TA) to 2-hydroxyterephthalic acid (2-HTA) was used to detect
hydroxyl radicals (•OH). The products of these reactions are easily measurable by
spectrometric and fluorimetric techniques [199]. TiO2 NP suspensions were prepared via
probe sonication, and the inorganic species (carbonate, chloride, nitrate, phosphate and
sulfate) were dosed as potassium salts at several concentrations. TiO2 suspensions were
photoactivated by UV-A irradiation, and the reactive species concentrations were
evaluated over time along with TiO2 aggregate characterization. This research has strong
implications not only for TiO2 photocatalysis engineering but also for the assessment of
the risk related to the environmental release of engineered nanomaterial.
4.2.2 Materials and Methods
4.2.2.1 Reagent Solutions
KOH, KI, I2, KHCO3, KCl, KNO3, KH2PO4, K2HPO4 and K2SO4 were purchased
from VWR International (USA). Stock solutions of inorganic anions (0.1 M) were
91
prepared in deionized water for carbonate. The starch solution was prepared by adding
0.5 g of starch (VWR International) in 100 mL of boiling deionized water and mixing
until complete dissolution. After 12 h settling, the supernatant was collected. TA (0.5
mM) and 2-HTA (0.125 mM) solutions were prepared by dosing 16.72 mg of TA (Sigma-
Aldrich, USA) and 4.69 mg of 2-HTA (Sigma-Aldrich) in 200 mL of deionized water,
adjusting pH to 7.9 with KOH, and mixing on a magnetic stir plate overnight.
4.2.2.2 TiO2 Dispersion and Suspension Characterization
Experiments were performed with P25 Aeroxide TiO2 nanoparticles (NPs)
(Evonik, Germany). TiO2 stock suspensions (40 mg L-1) were prepared by adding 2 mg of
P25 TiO2 to 50 mL of deionized water and stock solutions of inorganic anions,
depending on the test conditions. Following the protocol of Taurozzi et al., NPs were
dispersed via probe sonication (Q700, QSonica, USA) for 6 minutes in pulse mode (12 s
ON / 3 s OFF) with 1/2” diameter tip in the presence of a small concentration of
inorganic anions (1 mM) [178]. pH was adjusted to 7.9 with KOH (0.1 M) after
sonication, and the remaining inorganic anions were added to the desired concentration
afterwards when stocks were diluted in the reaction beaker.
Laser diffraction (Mastersizer 3000, Malvern Instruments, England) was used to
determine TiO2 aggregate size and fractal dimension. In-line light scattering
measurements were performed using a peristaltic pump located downstream of the
instrument by which samples were fed from a beaker continuously mixed on a magnetic
92
stirrer at 400 RPM. Time resolved size measurements, reported in terms of median
diameter (D50) both as number weighted and volume weighted intensity, were taken
every 2 minutes for 30 minutes. Meanwhile, static light scattering (SLS), consisting of
measurements of the scattering intensity, I(q), over a range of scattering vectors, q,
corresponding to 0.8 - 42° for each sample were recorded simultaneously by an array of
detectors. From Equation 28, the aggregate fractal dimension, Df, can be determined as
the negative slope of the linear power law region when plotting log(I(q)) vs log(q) [215]:
I(q) = q−Df (Eq 28.)
The multiple detectors of the Mastersizer correspond to scattering vectors from
0.18 – 9.47 µm-1, with q related to the detector angle, theta, as follows
q = 4πnλsinθ2
(Eq 29.)
where n is the refractive index of the medium (1.33 for water) and λ is the wavelength of
scattered light (633 nm) [194, 215]. Equation 28 holds when
qRh >1 (Eq 30.)
i.e. for length scales, (q-1) less than the radius of the aggregate, Rh.
TiO2 nanoparticle pH titrations were performed on 50 mL suspensions of 20 mg
L-1 P25 TiO2 at various anion concentrations. Suspensions were continually stirred while
pH was monitored and adjusted with either 0.01 M or 0.1 M KOH or HCl. At each pH
step, electrophoretic mobility of the nanoparticles was measured (Zetasizer Nano ZS,
Malvern Instruments, England). Zeta potential was then calculated via the Henry
93
equation. The isoelectric point (IEP) was identified as the pH for which the EPM of a
suspension was zero.
4.2.2.3 TiO2 Nanoparticle Reactivity Measurement
NP exposure to UV-A was performed in a box equipped with two 15 W
fluorescent UV lamps (TL-D 15W BLB SLV, Philips, Netherlands) held at 22±1 °C
through the use of a chiller. Samples were placed in 10 mL reaction beakers and
continuously mixed on a magnetic stirrer at 400 RPM. The radiation intensity at the
liquid upper surface was 1.9 mW cm-2, centered at 365 nm (see Appendix B) and
monitored by means of a ILT1400 radiometer equipped with a SEL033 UV-A filter
(International Light Technologies, USA). The geometry of the reaction beaker (H=3.5 cm,
Ø=5 cm) produced a 0.5 cm thickness liquid layer, including the stir bar (3 x 10 mm).
4.2.2.4 Photogenerated Holes
To evaluate the production of photogenerated holes in the presence of the
inorganic anions, reaction beakers containing 20 mg L-1 P25 TiO2, 50 mM KI and various
concentrations of anion (0, 0.5, 1, 2.5, 5, 12.5 mM) were exposed to UV light. Solutions
were irradiated for a total of 30 minutes, and all tests were repeated four times. Every 5
minutes a 0.5 mL aliquot was added to a solution of starch in a 1:1 volume ratio. The
resulting absorbance at 585 nm was then measured in a plastic cuvette (10 mm optical
path) vs. deionized water by spectrophotometer (Cary 100, Agilent, USA). Absorbance
values were related to iodine concentration by a standard curve, which was evaluated
94
by spectrometric measures on non-irradiated samples over the range of 0.0125 to 0.125
mM iodine after subtracting the TiO2 suspension contribution. The concentration of
photogenerated holes was stoichiometrically estimated as twice the produced iodine
concentration, according to the following reaction:
2I − + 2h+ → I2 (Eq 31.)
4.2.2.5 Hydroxyl Radicals
Hydroxyl radical production tests were performed on samples consisting of 20
mg L-1 P25 TiO2, 0.125 mM TA, and various concentrations of inorganic anions (0, 0.5,
2.5, 5, 12.5, 25 mM of carbonate, chloride, nitrate, phosphate and sulfate) and irradiated
for 30 minutes. All tests were repeated four times. Every 5 minutes a 0.5 mL aliquot was
added to 0.75 mL deionized water in a 2 mL centrifuge tube. Samples were then
centrifuged for 5 minutes at 12,000 RPM to separate the TiO2 from the solution. 1 mL of
the resulting supernatant was transferred to a plastic cuvette (10 mm optical path) and
fluorescence was measured (Varian Eclipse fluorometer, Agilent, USA) at an excitation
wavelength set to 315 nm and emission maximum at 425 nm. Fluorescence was related
to the concentration of 2-HTA by a standard curve by diluting a 0.125 mM 2-HTA stock
solution to obtain a range from 0.0025 - 0.1 M, adopting the same procedure as for the
samples. Hydroxyl radical concentrations were estimated by the following reaction,
assuming 80% trapping efficiency for OH• by TA [87] and considering sample dilution in
deionised water:
95
TA+ •OH→ 2−HTA (Eq 32.)
4.2.3 Results and Discussion
The strong influence of inorganic anions on TiO2 nanoparticle photocatalysis,
resulting from the combination of chemical-physical phenomena is presented here first
in terms of impacts on NP aggregation (Section 4.3.1). The anion influence on TiO2
reactivity is then discussed in terms of photogenerated holes (Section 4.3.2) and
hydroxyl radicals (Section 4.3.3).
4.2.3.1 TiO2 Aggregate Characterization
Sonicating TiO2 nanoparticles in deionized water created aggregates with a stable
number weighted D50 of 0.095±0.004 µm and volume weighted D50 increasing from
0.591±0.296 µm to 6.84±7.63 µm over 30 minutes (Data in Appendix B). The difference in
values arising from the weighting methods is due to the polydispersity of TiO2
suspensions, indicating that the vast majority of aggregates exist at the smaller size class
with relatively few large aggregates influencing the volume weighted D50. This smaller
fraction is expected to be the most photoactive, as non-ROS producing pathways (e.g.
quenching, recombination) increase in frequency with aggregation [219]. Suspensions in
DI were repeated five times; the large standard deviation associated with the volume
weighted D50 arises from the instability of the TiO2 aggregates in the mixed solution.
96
4.2.3.2 Aggregation in Inorganic Salts
Monitoring TiO2 D50 in the presence of inorganic salts yielded two contrasting
trends (i) anions that did not limit NP destabilization, and (ii) anions that stabilized NP
suspensions through specific surface interactions. Figure 25 shows TiO2 aggregation
over time for each anion at 0.5 mM and 25 mM. Aggregates in chloride and nitrate
aggregated significantly over 30 minutes at all concentrations, increasing from an initial
TiO2 number weighted D50 similar to that of aggregates suspended in DI to a final size of
a micron or more (Data in Appendix B). On the other hand, phosphate, carbonate, and
sulfate all had a stabilizing effect. The greatest effect was seen for phosphate for which
no increase in aggregate size was observed, even at concentrations as high as 25 mM.
Some aggregation proceeded for suspensions containing carbonate at 12.5 and 25 mM.
The addition of sulfate in solution did result in a decrease in aggregation, though this
was much less pronounced than for either phosphate or carbonate, and did not exhibit
concentration dependence over the range examined.
97
Figure)1)
0 5 10 15 20 25 30−10
0
10
20
30
40
50
60
time (min)
D50
(µm
)
PhosphateChlorideNitrateCarbonateSulfate
(a))0.5mM) (b))25mM)
0 5 10 15 20 25 30−10
0
10
20
30
40
50
60
time (min)
D50
(µm
)
PhosphateChlorideNitrateCarbonateSulfate
(b))(a))
Figure 22: TiO2 aggregation over time for each anion at (a) 0.5 mM and (b) 25 mM. Data shown is volume weighted D50 (mean±std)
Looking at the initial rate of aggregation for chloride (Figure 23) and nitrate (data
in Appendix B), no clear influence of ionic strength can be discerned, although the
attachment efficiency, α, is generally expected increase with salt concentration until it
reaches unity when the critical coagulation concentration is reached. Calculations using
the Gouy Chapman model indicate that attractive forces dominate (i.e. α = 1) for 12.5
and 25 mM potassium chloride (data in Appendix B). For the lower three concentrations
(0.5, 2.5, and 5 mM) however, the electrical double layer remains large enough that
electrostatic repulsion should hinder aggregation. These results agree well with
Solovitch et al. who determined the CCC for 50 mg L-1 anatase at pH 8 in unmixed
experiments to be between 10 and 40 mM NaCl [303]. That no difference was observed
in initial aggregation rates here, despite the differences in particle stability, highlights
the strong influence of mixing on aggregation rate for these conditions. The initial
aggregation reached a plateau after approximately 14 minutes; presumably this is the
98
point at which breakup and reordering due to shear becomes significant. Similar
behavior was observed by Gardner et al. in studying the aggregation kinetics of
hematite particles in a Couette device [213]. For suspensions to which carbonate was
added, the initial rate of aggregation was greater for samples at 25 mM than for 12.5 mM
of concentration, as would be expected under DLVO theory, though the aggregate size
for both quickly stabilizes after which no further aggregation is observed.
Figure)2)
Chloride,)vol)
0 5 10 15 20 25 300
10
20
30
40
50
60
70
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
Figure 23: TiO2 aggregation (volume weighted D50 mean±std) vs. time as a function of chloride concentration.
4.2.3.3 Fractal Dimension
SLS measurements were collected for each aggregation experiment. At early time
points, the size of aggregates is small compared to the length scales probed, and plots of
log(I) vs log(q) were not linear over a large enough range of the scattering vectors
considered (0.110 µm to 5.62 µm) (Data in Appendix B). Thus, the constraint of Equation
99
30 is not met and related light scattering data were not sufficient to calculate Df.
Furthermore, as no aggregation was observed for any phosphate concentration and for
low carbonate concentrations (<12.5 mM), fractal dimensions were not obtainable. For
suspensions in which aggregation was observed, the average Df values over the final
three measurements (26-30 minutes) are shown in Figure 24. Broadly, Df values were
lower for suspensions that were observed to quickly aggregate (nitrate, 2.33±0.001;
chloride, 2.32±0.002; and sulfate, 2.31±0.002 at 25 mM), indicative of more open
Figure)3)
12.5mM 25mM2
2.1
2.2
2.3
2.4
2.5
2.6
Df
CarbonateNitrateChlorideSulfate
0.5mM 2.5mM 5mM 12.5mM 25mM2
2.1
2.2
2.3
2.4
2.5
2.6
Df
CarbonateNitrateChlorideSulfate
USE)ONLY)(b)?)
(b))(a))
Figure 24: Fractal dimension of TiO2 aggregates for all anions at 12.5 and 25 mM. Df values are the average of the final 3 measurements (mean±std).
aggregates. Conversely, the higher fractal value for carbonate (2.40±0.02 at 25 mM)
agrees well with the slower observed aggregation that would result in smaller, denser
aggregates. The concentration of anion in solution did not significantly impact aggregate
100
structure. For nitrate, chloride, and sulfate, values of Df are observed to center around
2.3 for all concentrations and fall within the range of what is considered reaction-limited
aggregation, indicative of stable particles. This likely arises from shear due to mixing,
resulting in aggregate breakup and reordering, leading to denser clusters. This lack of
concentration dependence also agrees well with initial rates of aggregation discussed
previously.
4.2.3.4 Aggregate Stability
To identify if the various anions impacted the surface chemistry of TiO2
aggregates, NP suspensions were titrated and the IEP for each suspension was identified
(Data in Appendix B). EPM vs pH was measured in the range of pH 3-9. As can be seen
in Figure 28, the addition of chloride and nitrate, generally considered indifferent
electrolytes, did not significantly impact the IEP (6.6±0.1 versus 6±1 [304]), while
carbonate and phosphate interacted strongly with the surface, resulting in large shifts in
the IEP, the extent of which varied with anion concentration. Sulfate depressed the IEP
from that of chloride and nitrate to 5.6±0.2 but did not vary greatly with concentration.
The impact of carbonate, however, was drastic. At 0.005 and 0.05 mM, the IEP mirrored
that of sulfate, though increasing the concentration decreased the IEP to a low of 2.8 at 5
mM. The IEP then rebounded with a further increase in carbonate concentration.
Phosphate proved to even more strongly influence the IEP as the concentration
increased from 0.005 mM to 0.5 mM, at which point NPs were negatively charged for the
101
entire range of pH measured. Further increasing phosphate concentration resulted in the
slight increase in IEP, though this was not as pronounced a recovery as was observed
with carbonate. Titrations above 12.5 mM were not performed as the ionic strength
interfered with EPM data quality.
2 3 4 5 6 7 8 9 10−50
−40
−30
−20
−10
0
10
20
30
40
50
pH
ZP (m
V)
PhosphateNitrateChlorideCarbonateSulfate
Figure)4)
10−3 10−2 10−1 100 101 1022
3
4
5
6
7
8
[Anion] (mM)
IEP
(pH
)
Phosphate Carbonate Sulfate Nitrate Chloride
All)[anion])=)0.5mM)
(b))(a))
Figure 25: (a) Zeta potential vs. pH for all anions at 0.5 mM. (b) IEP vs. concentration for all anions. IEP for 0.5 mM phosphate was not determined as the ζ
potential was negative over the range of pH tested.
That carbonate, phosphate, and sulfate were able to influence the IEP speaks to
their ability to participate in inner sphere ligand exchange with the surface oxides and
hydroxides present on TiO2 [305-307]. The magnitude of anion influence increases with
surface coverage and depends on the dissociation and binding constants of the species
present [159, 227, 285]. The increase in IEP observed for phosphate and carbonate at
concentrations of 5mM and higher may be due to changes in ligand complexation. Inner
sphere complexes can form mono- bi- or tridentate complexes depending on ligand type,
pH, lattice structure and surface defects [308-311]. These complexes are associated with
different binding energies and associated protonation pKa values [40, 227, 312]. As
102
concentration increases and the more energetically favorable binding sites become filled,
adsorption will transition to less favorable sites. Phosphate has been shown to favor
bidentate complex formation with the surface of TiO2, which becomes protonated at low
pH. Monodentate adsorption results in a protonated complex that becomes doubly
protonated at low pH [227]. The addition of sulfate quickly reaches a maximum impact
beyond which the further introduction in solution has little to no effect, suggesting that
the interactions of sulfate with the surface are less significant than phosphate and
carbonate. These data fit well with observations of aggregate size (Section 3.1.2), in
which TiO2 suspended in phosphate and low concentrations of carbonate do not induce
aggregation over the course of 30 minutes, in contrast to nitrate and chloride. Thus, the
picture that arises in one in which the anions participating in specific surface
interactions with TiO2 strongly impact the stability of the NPs and the size of the
aggregates formed in solution.
4.2.3.5 Influence of ROS Detection Method
KI was used as the probe compound for photogenerated hole oxidation. The
addition of KI resulted in NP aggregation from an initial size similar to NPs in DI
(0.099±0.004 µm, number weighted D50) to 1.23±0.56 µm over 30 minutes (0.547±0.105
µm increasing to 31.13±8.23 µm, volume weighted D50) (Data in Appendix B). The initial
rate of this aggregation was indistinguishable from that observed for nitrate and
chloride, discussed previously (Data in Appendix B).
103
Size data, both as number weighted and volume weighted D50, over time for TiO2
suspensions at various iodide concentrations (0, 50 mM) and relative size distributions
are reported in Appendix B. While the probe itself induces aggregation, it does not do so
in a manner that is appreciably different from the addition of the indifferent electrolytes.
No significant aggregation was observed in the presence of terephthalic acid, thus
aggregate size data were identical to that for DI.
4.2.3.6 TiO2 Photoreactivity – Photogenerated Holes
The influence of inorganic anions on the production of photogenerated holes was
assessed by via the oxidation of iodide. The initial KI concentration was determined by
measuring the response at various concentrations using suspensions of TiO2 in DI water
(data in Appendix B) [199, 313]. A saturation trend in oxidation was observed, and 50
mM was selected as the optimal concentration. The high concentration of iodide
required to reach saturation likely arises from the negative surface charge of TiO2 at pH
7.9, both electrostatically limiting iodide transport to the aggregate surface and resulting
in an increased concentration of hydroxide ions present in solution and at the NP
surface.
Upon addition of the various anions, a change in the rate of iodide oxidation was
only observed in the presence of carbonate and phosphate. Chloride, nitrate and sulfate
at concentrations between 0.5 and 12.5 mM did not cause any significant change in
iodide oxidation from samples containing only KI. Experimental results after 30 minutes
104
of irradiation at various carbonate and phosphate concentrations are shown in Figure 26.
Experimental data for chloride, nitrate, and sulfate were not significantly different from
results obtained for TiO2 suspensions without inorganic ions and are not shown.
Oxidized iodide concentrations for these three anions overlap the results when only KI
is present in solution (data in Appendix B). The lack of statistically significant
differences between series with or without anions for chloride, nitrate and sulfate was
verified by means of Student’s t-test (p-value < 0.005).
Fig)5)alternaVve)
0 5 10 150.005
0.01
0.015
0.02
0.025
0.03
[Anion] (mM)
I ox (m
M)
CarbonatePhosphate
Figure 26: Oxidized iodide concentration (mean±std) versus anion concentration after 30 minute irradiation (50 mM KI) as a function of species in
solution (carbonate, phosphate). The blue dotted line is the iodide concentration of TiO2 suspensions containing only 50 mM KI.
The addition of carbonate and phosphate increased the rate of photogenerated
hole oxidation for phosphate at concentrations up to 5 mM and for all carbonate
concentrations tested as compared to solutions containing only the probe. The largest
105
increase was observed at 0.5 mM and was the same as that for carbonate and phosphate
anions, after which increasing the concentration reduced hole oxidation to a level
observed in DI suspensions. These findings agree with the trend of aggregate
stabilization reported in Section 4.2.3.2, which would result in increased reactivity, and
suggests that at low concentrations, little negative effect on reactivity due to specific
surface interactions would be expected. The reduction in reactivity observed with
increasing anion concentration is attributed to the presence of phosphate or carbonate
occupying surface sites and preventing iodide from interacting with the photogenerated
holes. The effect of phosphate addition was greater than carbonate, despite observed
aggregation of carbonate at 12.5 mM, indicating that the decrease in photogenerated
holes is primarily due to the greater surface interactions of phosphate and TiO2.
The interactions between anions and TiO2 nanoparticles has been described by
Chen et al., as a Langmuir type adsorption model [286]. The oxidation of iodide can thus
be modeled as a Langmuir-Hinshelwood (LH) mechanism, even when considering the
presence of other anions in solution, per Equation 33 [313].
r = d[I− ]
dt= −Krθ I− = −KR
KA,I[I− ]
1+KA,I[I− ]+ KA,Anion[Anion]
Anion∑
(Eq 33.)
Here r is the reaction rate, θI– is the surface coverage by iodide, [I–] is the
concentration of iodide, KA,I is the equilibrium constant of adsorption for iodide, KA,Anion
is the equilibrium constant of adsorption for a given anion and KR is the LH reaction
106
constant for iodide oxidation. Estimates of KA,I and KR were calculated from the
experimental data reported in Appendix B in the absence of other anionic species
following the procedure employed by Herrmann and Pichat (see Appendix B), the
results of which are shown in
Table 3 [313]. The high R2 values indicate that the LH mechanism is appropriate for
describing iodide oxidation. These values are comparable with those obtained
previously for iodide oxidation in non-sonicated suspensions of TiO2 at 40 mg L-1 [199].
Table 3: Estimated values of the equilibrium adsorption coefficient, KA, for iodide, carbonate and phosphate and estimated values of LH reaction constants for
iodide oxidation, KR. R2 values are the coefficients of determination.
Ionic species KA (M-1) KR (M s-1) R2 Iodide 13.2 2.24E-08 0.988 Carbonate 142.7 4.16E-08 1.000 Phosphate 251.8 4.02E-08 0.954
Similarly, the LH mechanism was used to model the effect of inorganic anions on
the formation of photogenerated holes. Here it is essential to take into account the
competition for the active sites in LH mechanism among inorganic anions. A procedure
for KA,Anion and KR estimation in the presence of anionic competition for surface
adsorption is described in Appendix B, in which the equilibrium constant of adsorption
for iodide was taken from the previous calculation in the absence of other inorganic
anions. The LH mechanism fits the experimental data well (R2>0.950). Larger values of
107
KA,Anion indicate greater active site occupation by a given anion compared to others at the
same concentration. Hence, iodide, having the smallest KA,Anion value, is the least favored
at the TiO2 surface under the present operating conditions. KA,Anion for phosphate was
much higher than for carbonate and indicates a greater affinity for the TiO2 surface
which serves to more strongly block iodine from the surface and reduce hole reactivity.
KA,Anion for nitrate, chloride, and sulfate were not calculated, as no adsorption in the
presence of iodide was observed. The influence of the associated cation, i.e. potassium,
on reactivity was considered negligible, in agreement with previous works [314].
4.2.3.7 TiO2 Photoreactivity - Hydroxyl Radicals
The generation of hydroxyl radicals was evaluated by monitoring the
fluorescence of 2-HTA as a function of inorganic anion concentration. TA is a selective
hydroxyl radical probe, whose reaction occurs via a single step hydroxylation process in
the absence of surface adsorption [76]. At pH 7.9, the carboxylic acid groups on TA are
deprotonated and the molecule is electrostatically repulsed from the negatively charged
TiO2 surface.
Table 4: Rate constants of hydroxyl radicals and inorganic anions.
Anionic species Value (M-1 s-1) Reference HCO3- 8.5x106 [70]
Cl- 4.3x109 [70] NO3- 1.4x108 [315]
HPO42- 6x105 [316] SO42- 1.0x1010 [317]
108
Concentrations of 2-HTA for various anionic species after 30 minutes of
irradiation are shown in Figure 27 (a), while relative time series are reported in Figure 27
(b). For all anions, increasing the concentration in solution decreased the amount of ROS
detected, as would be expected for potential hydroxyl radical quenchers. However, rate
constants for the reaction of hydroxyl radicals with each inorganic anion, available from
literature and reported in Table 4, do not provide a comprehensive explanation for this,
as the expected impact on �OH detection differs from experimental results. For example,
the given reactivity of sulfate with hydroxyl radicals (1x1010 M-1 s-1) is four orders of
magnitude greater than carbonate (8.5x106 M-1 s-1), although in this study carbonate
produced the greatest decrease in �OH detection and sulfate produced the least.
While phosphate and carbonate have the smallest reaction rates, their
introduction decreased total hydroxyl radical production to the greatest extent. This is
best explained by considering the conditions that lead to hydroxyl radical generation,
i.e. hole production and aggregate size. The specific interactions between the TiO2
surface and carbonate and phosphate were shown to initially increase photogenerated
holes, and for carbonate this is still seen in hydroxyl radical generation, however this
effect is quickly counteracted by scavenging when increasing the carbonate
concentration. Phosphate produces a strong initial decrease in �OH generation at 0.5mM
despite the observed aggregate stabilization and increase in hole production, after which
the impact of additional anion is relatively small. Indeed, by 12.5 mM the quenching
109
�OH detection for carbonate is equal that of phosphate. By 25 mM, the addition of
carbonate decreased �OH production by c.a. 70%. This greater effect of carbonate than
phosphate, contrary to what is observed for hole oxidation is likely due to the order of
magnitude increase in k*OH for carbonate. Furthermore, the strong impact of both anions,
given their relative inactivity with �OH compared to sulfate, chloride, and nitrate
(greater than two orders of magnitude smaller reaction rate), is attributable to their
being drawn to and interacting with the TiO2 surface. This not only limits the number of
surface hydroxides capable of reacting with the photogenerated hole to produce �OH,
but it also increases the local concentration of quenching agents (carbonate and
phosphate) near the NP surface relative to tests using the same concentrations of the
other anions, thus decreasing both the lifetime of generated hydroxyl radicals.
Figure)6)
0 5 10 15 20 25 300.004
0.006
0.008
0.01
0.012
0.014
0.016
0.018
[Anion] (mM)
[2−H
TA] (
mM
)
CarbonateChlorideNitratePhosphateSulfate
0 5 10 15 20 25 300.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
[Anion] (mM)
Nor
mal
ized
Hyd
roxy
l Rad
ical
Gen
erat
ion
CarbonateChlorideNitratePhosphateSulfate
(b))(a))
Figure 27: (a) 2-Hydroxyterephthalic acid concentration (mean±std) vs. anion concentration after 30 minute irradiation as a function of species in solution. (b)
Hydroxyl radical generation normalized to 0.5 mM anion concentration (mean±std).
The trend for nitrate, chloride, and sulfate is more difficult to discern. A large
initial decrease in �OH at 0.5 mM compared to DI is observed for sulfate and nitrate,
110
whereas chloride initially has little discernable impact. Surprisingly, the addition of up
to 25 mM sulfate only decreased �OH detection by c.a. 3%. Since the interaction between
TiO2 surface and TA is negligible [76] and TiO2 aggregate characteristics are similar, the
negative influence of nitrate and chloride on terephthalic acid oxidation can be
considered the result of the interplay of two primary phenomena: (i) a reduction in
hydroxyl radical generation due to the occupation of active sites by anionic species
instead of hydroxyl ions, (ii) the direct quenching of hydroxyl radicals by inorganic
anions. The absorption of incoming photons by inorganic anions resulting in reduced
NP photoactivation, is considered negligible [16]. The poor description given by the LH
mechanism (R2 < 0.7) supports this assertion that the impact on �OH is due to a
combination of phenomena, where mass transfer and electrostatic interactions possibly
play a relevant role.
4.2.4 Conclusions
In this work, TiO2 reactivity has been explored both at the NP surface
(photogenerated hole oxidation) and in the bulk solution (�OH detection). Only
carbonate and phosphate inhibited the oxidation of iodide through specific surface
interactions. Chloride, nitrate and sulfate either do not interact with TiO2 or create
relatively weak bonds (with respect to carbonate, phosphate) on the TiO2 surface that do
not hinder iodide. For sulfate, it would appear that while the anion does specifically
interact with the surface, these interactions are not sufficient to limit iodide from
111
reaching and reacting with the surface. The lack of aggregate stabilization and the
measured IEPs corresponding to bare TiO2 suggest that for nitrate and chloride, no
specific interactions take place. The response of �OH reactivity to the addition of anions
is more complicated and arises from the multiple existing pathways by which the NPs
can be impacted. The overall influence on bulk �OH concentration was observed to
follow the trend carbonate > phosphate > nitrate > chloride ≈ sulfate.
Previous studies looking at the impact of anions on TiO2 photocatalysis have
monitored the overall degradation of pollutants using the LH mechanism, without
differentiating direct hole oxidation from hydroxyl radicals [285, 291]. Others, such Chen
et al. investigating the degradation of dichloroethylene, assumed adsorption as a
necessary precursor to degradation, but ruled out the likelihood of direct hole oxidation
[286]. The intensity of inorganic anion influence on TiO2 reactivity differs in this study
from what has been reported, among which, however, there is no general consensus.
Differences between studies seem to be strongly related to experimental conditions,
namely solution pH, degradation target chemistry, and TiO2 characteristics (e.g.
aggregate size, dispersed versus fixed catalyst), affecting the balance among the
involved chemical-physical processes. As a result, generalization between studies is
difficult. Often, little attempt has been made to discern surface reactivity from bulk
oxidation, however as is evidenced here, the anion impact on hole generation can be
different than the impact on ROS generation. Thus, the mechanism by which a
112
compound or organism is impacted will play a large role in determining the effect of
changes in reactivity.
This work highlights the fact that the reactivity of NPs will be strongly
influenced by the waters they are released into, due to the interplay of several
phsyicochemical phenomena. Inorganic anion adsorption on the TiO2 surface happens
via a Langmuirian model and creates a competition for active sites with other
compounds in solution, both contaminants and hydroxyl ions. The effect of this
adsorption can consist of: (i) stabilizing NPs, (ii) hindering the interaction of pollutants,
(iii) suppressing hydroxyl radical generation by substituting hydroxyl ions. Some waters
containing chloride and nitrate may have little impact on reactivity but will reduce NP
transport via aggregation, while waters containing even low levels of phosphate and
carbonate may decrease “acute” reactivity but stabilize NPs such that their lifetime in
the water column is increased. The ultimate impact of NP photoreactivity in
environmental waters (risk) will be the combination of reactivity (hazard) and stability
or transport (exposure).
113
4.3 Chlorpyrifos Degradation via Photoreactive TiO2 Nanoparticles: Assessing the Impact of a Multi-Component Degradation Scenario
ABSTRACT
High concentrations of pesticides enter surface waters following agricultural
application, raising environmental and human health concerns. Bioremediation, while
an attractive strategy for mitigating the risk associated with contaminated waters, can
require can require extensive periods of time. Furthermore, high concentrations of
pollutants may be toxic to or limit the efficacy of bacteria. Alternatively, the use of
photoreactive nanoparticles has shown promise for contaminant degradation and
surface water remediation. However, it remains uncertain how the complexity of natural
waters will impact the photodegradation process. Here, we investigate the
photoreactivity of titanium dioxide nanoparticles, the capability to degrade the pesticide
chlorpyrifos (CPF), and the effect of and impact on bacteria during the
photodegradation process. Loss of chlorpyrifos in solution resulted solely from
photocatalytic oxidation. Degradation of chlorpyrifos to chlorpyrifos oxon (CPF Oxon)
and 3,5,6-trichloro-2-pyridinol (TCP) was observed and effectively modeled for the
given reactor conditions. The relative affinity of bacteria and chlorpyrifos for the
nanoparticle surface decreased the amount of ROS present in the bulk, measured using a
hydrophilic probe, indicating that measurements in simplified systems are likely to
overestimate reactivity in complex environments.
114
4.3.1 Introduction
Chlorpyrifos (CPF) is an organophosphate pesticide used for agricultural
applications such as cotton, corn, and fruit trees, as well as at high concentrations in
wood for anti-termite applications [318-321]. As a result, CPF has been frequently
observed in agricultural runoff and urban streams [322-325]. Both CPF and its primary
oxidation products, chlorpyrifos oxon (CPF Oxon) and 3,5,6-trichloropyridinol (TCP),
are of considerable environmental concern due to high fish and aquatic invertebrate
toxicity, as well as known human health concerns [322, 326-334].
CPF was designed to be susceptible to base catalysis, with a half-life in water on
the order of days, though studies of adsorbed CPF indicate that the half-life can range
from hundreds of days to years [321, 326, 335-337]. Gebremariam et al., in reviewing the
literature, observed a significant spread in the extent of CPF adsorption to various soil
types, though a strong linear relationship was seen with soil organic content, implicating
the hydrophobic binding sites present on organic matter [321]. Rose et al. observed that
the loss of CPF from a water column in the presence of natural colloidal matter occurred
more rapidly than other common pesticides studied [338]. However, rainfall has been
shown to be sufficient for CPF-particle resuspension, which may lead to increased
bioavailability and transport [321, 326, 327].
Photocatalytic degradation holds promise for dealing with aquatic contamination
[14, 15, 339-343]. Nanoscale TiO2 is widely studied for remediation applications due to
115
its photochemical properties and ability to provide a steady, continual source of reactive
oxygen species (ROS) [344, 345]. CPF degradation experiments employing TiO2 as the
radical generator often employ high concentrations of the catalyst (0.1 – 12 g L-1) in an
effort to maximize radical production and minimize treatment time, with CPF loss often
occurring within hours [344, 346, 347]. What remains unclear, however, is to what extent
the observed loss of CPF from the solution is due to sorption versus degradation. CPF’s
affinity for soils may not translate to TiO2, as sorption of CPF increases with organic
content [321]. Furthermore, little absorption has been seen for similar pesticides when
employing metal oxide nanoparticles (NPs) [337, 348-350]. This question is of particular
importance in light of the high surface area of NPs and their potential to act as pollutant
transport vectors should incomplete degradation occur [351-353].
Additionally, many types of bacteria have been observed to degrade CPF,
including Bacillus and Acinetobacter strains isolated from groundwater sources, either
using the pesticide as the sole carbon source or cometabolically [319, 336, 354-358].
Biodegradation may be an appealing approach to CPF remediation, though longer
residence times are necessary compared to chemical or external remediation techniques
[359].
High levels of pollutants may also be toxic to bacteria and hinder bioremediation
efforts. Toxicity studies of CPF and its degradation products have shown that there is an
impact on both individual species and entire bacterial communities [360-362]. When
116
concentrations of CPF are high, pretreatment of contaminated water may improve
remediation efforts. TCP was calculated to be 2.5 times less toxic than CPF for a model
bacterium, indicating that pretreatment achieving even partial degradation of a
pollutant may facilitate bioremediation efforts [322].
A mixed or sequential approach combining nanoremediation with a biological
component may be a more efficient and cost effective treatment method. Within this
context, our understanding of the microbial response in a UV/TiO2 remediation system
is currently lacking, with the concern existing that any nanoremediation strategy will
harm the natural bacterial community. ROS production by the UV/TiO2 system will
possibly be of more consequence than the pollutant [146, 222, 360, 363-365].
In this paper we investigate the process of CPF loss in water utilizing UV/TiO2,
hypothesizing that CPF loss in the system is primarily due to photocatalytic
degradation, rather than sorption. Additionally, we test the hypothesis that the UV/TiO2
process will inactivate bacteria to a greater extent than will CPF or its degradation
products, reducing the likelihood of bioremediation.
4.3.2 Materials and Methods
4.3.2.1 TiO2 Suspension and Characterization
Experiments were performed with P25 Aeroxide TiO2 (Evonik Industries, Essen,
Germany), a nanoparticulate TiO2 consisting of fused rutile and anatase crystalline
phases at roughly a 20/80% ratio, respectively. TiO2 NPs were suspended in a Minimal
117
Davis media (MD) designed to stabilize the particles, buffer the pH at 7.5, and maintain
bacterial viability, as described by Lyon et al. [366] Suspensions were produced via
probe sonication (Q700, QSonica, Newton, CT, USA) run in pulse mode (12s on, 3s off)
for 6 minutes total, following the protocol from Taurozzi et al., and subsequently
characterized in terms of aggregate size, surface charge, and ROS production [178].
Dynamic light scattering (ALV-CGS3, ALV-GMBH, Langen, Germany) and transmission
electron microscopy (TEM) (Tecnai G² Twin, FEI, Hillsboro, OR, USA) were performed
to determine aggregate size, while electrophoretic mobility measurements (EPM) were
taken with the Zetasizer Nano ZS (Malvern, Bedford, MA, USA).
4.3.2.2 UV Irradiation and Photoreactivity
CPF was exposed to TiO2 both in the presence and absence of UV light to
investigate the relative effects of sorption and degradation. UV-A irradiation was
performed with two 15 W fluorescent bulbs (Philips TLD 15W BLB, Koninklijke Philips,
Eindhoven, The Netherlands) having a peak emission at 365 ± 15nm. Samples were
placed in batch reactors and continually stirred using glass stir bars via magnetic stir
plates with UV irradiation from above. For UV experiments involving CPF, reactors
were covered with glass petri dishes to minimize loss due to volatilization. The
temperature inside the reaction chamber was kept at 23oC. The net irradiance at the
suspension surface was 2.3 mW cm-2 for bacterial inactivation experiments and 2.0 mW
cm-2 for CPF degradation tests, measured by a UVX radiometer with UV-A filter (UVP,
118
inc., Upland, CA, USA). TiO2 reactivity was also characterized with and without CPF to
explore how CPF affinity for the TiO2 surface relative to other species in water impacts
degradation efficacy.
ROS production was measured via probe compounds and quenching agents. For
hydroxyl radical (�OH) measurements, terephthalic acid (TA) (98% Sigma-Aldrich, St.
Louis, MO, USA) and N-acetyl-L-cysteine (N-AC) (Sigma-Aldrich) were used as probe
and quencher, respectively, where �OH generation was measured as the increase in
fluorescent units (FSU) resulting from the oxidation of 125 µM TA to 2-
hydroxyterephthalic acid (2-HTA) (ex. 315 nm / em. 425 nm). Superoxide (O2�-) was
detected as the increase in FSU at 586 nm (excitation at 510 nm) resulting from the
transformation of 50 µM dihydroethidium (DHE) in solution to 2-hydroxyethidium (2-
HE) (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA). Superoxide dismutase
(SOD) (Sigma-Aldrich) was used as the O2�- quenching agent.
Dark controls were run to test CPF adsorption to the NP surface, CPF was spiked
into MD media containing TiO2 nanoparticles at 0ppm (control), 20 mg L-1, and 40 mg L-1
and allowed to mix in the dark. Reaction vials were covered tightly in aluminum foil
and were periodically sampled over 24 h. All concentrations were run in triplicate.
4.3.2.3 Bacterial Cultures
Gram Positive Bacillus subtilis strain 168 ATCC 23857 and Gram Negative
Acinetobacter baumannii ATCC 49466 (ATCC, Manassas, VA, USA) were chosen as model
119
bacteria based upon their environmental relevance, documented ability to biodegrade
CPF, and cell membrane structure. Pure cultures of A.baumannii and B.subtilis were
grown overnight at 37oC in tryptic soy broth. Cells were pelleted at 8,000 rpm for 10
minutes and washed thrice with MD. Cells were then resuspended in MD to a final
concentration of 109 cells mL-1. 2 mL aliquots of the bacteria suspension were added to
each batch reactor.
To determine if CPF degradation products impact bacteria, we have exposed
cultures of B.subtilis and A.baumannii to CPF, CPF oxon, and TCP, individually. Toxicity
of the products arising from photocatalytic degradation was investigated by exposing
CPF for 60 minutes to UV/TiO2. Suspensions of B. subtilis and A. baumannii were then
dosed with the resulting combination of CPF, TiO2 nanoparticles, and CPF degradation
products present in the system. Additionally, bacteria were exposed to the UV/TiO2 to
evaluate the biological impact of remediation attempts.
Samples taken from the reactors for inactivation measurements were serially
diluted to obtain a final concentration of 103 CFU mL-1. 100 μL of the final dilution was
then spread onto tryptic soy agar plates in triplicate and allowed to grow overnight at
37oC. The resulting colonies were then counted, and ROS and CPF toxicity was
calculated as log inactivation relative to time zero.
120
TEM images of Bacteria and TiO2 NPs were obtained by depositing samples on
lacey carbon grids. These were fixed with 2.5% glutaraldehyde, and sequentially
dehydrated and washed with graded ethanol and 1x PBS, respectively.
4.3.2.4 Chlorpyrifos
Chlorpyrifos (99.5%, Chem Service, West Chester, PA, USA) stocks were made in
methanol at 1.5 g L-1, which were diluted in MD to working stocks at 1.5 mg L-1. For all
experiments, CPF was spiked into samples at 375 µg L-1. Stocks of CPF degradation
products CPF Oxon (analytical standard, AccuStandard, New Haven, CT, USA) and
3,5,6-trichloro-2-pyridinol (TCP) (analytical standard, Sigma-Aldrich) suspensions were
similarly prepared.
4.3.2.5 Sample Preparation and Analysis
At each sampling point, a 0.2 mL aliquot was transferred to a 1 mL tapered base
glass reaction vial (Kimble Chase, Vineland, NJ) along with 10 μL formic acid to
aggregate TiO2 nanoparticles. Samples were centrifuged for 2 min at 9,000 rpm and the
supernatant transferred to an LC/MS vial and spiked with 100 μL each of 1 μgmL-1
deuterated chlorpyrifos and 13C6-3,5,6-trichloro-2-pyridinol (Cambridge Isotopes,
Cambridge, MA) and diluted to a final volume of 0.5 mL in 20% acetonitrile (ACN,
Honeywell Burdick & Jackson, Muskegon, MI) to match initial LC mobile phase
conditions (described below). Residue in the reaction vial was extracted three times with
300 μL ACN and 10 μL formic acid via sonication for 5 min followed by centrifugation
121
(described above). Extracts were combined, concentrated to 100 μL under N2, spiked
and diluted as described above. Glass pipettes were used at all times (e.g. when
sampling individual time points and transferring aliquots during extractions).
Table 5: Multiple reaction monitoring transitions used in the quantification of CPF and its degradation products.
Compound1 Description Transition2 Collision
Energy (V) S-Lens
(V) Polarity
D10-CPF Internal standard 359.9 > 98.9 (Q) 33 70 +
CPF Target analyte 349.9 > 96.9 (Q) 33 70 +
349.9 > 197.8 (c) 20 70 +
CPF Oxon Target analyte 334 > 277.8 (Q) 16 70 +
334 > 197.8 (c) 29 70 +
13C6-TCP Internal standard 2.9200 > 162.9 (Q) 19 60 -
TCP Target analyte 195.9 > 160.9 (Q) 18 60 -
195.9 > 35.1 (c) 35 60 -
1 Target analytes listed below an internal standard were quantified using that internal standard.
2 Transitions used for quantitation are designated "Q" in parentheses; transitions used for confirmation are designated "c" in parentheses.
CPF and its degradation products were analyzed by LC/MS-MS using electrospray
ionization in positive (CPF and CPF Oxon) and negative (TCP) modes with a Thermo
Accela UPLC system and Thermo TSQ Vantage mass spectrometer (Thermo Scientific).
Separation was achieved on a Synergi Polar RP column (50 mm x 2.0 mm, 2.5 µm
particle size, Phenomenex, San Jose, CA) and a 400 μL min-1 gradient of 5 mM formic
122
acid (A) and 5 mM formic acid (B) in ACN with the following conditions: 0-0.2 m 70:30
A:B to 1:99 A:B at 3-4 min with return to initial conditions at 5 min and re-equilibration
for 4 m. Source conditions included spray voltages +3.5 kV/-2.5 kV, vaporizer
temperature 350ºC, capillary temperature 375ºC, sheath gas 45 psi and auxiliary gas 10
psi. Target analytes and internal standards were monitored using the transitions
described in Table 5.
4.3.2.6 CAKE software package
The computer-assisted kinetic evaluation (CAKE) v3.1 software package,
developed by Tessella (Abingdon, UK) was used to determine degradation rates and
pathways. Single first order (SFO) fitting was employed for CPF degradation as well as
CPF Oxon and TCP formation. CPF degradation was assumed to proceed to CPF Oxon,
TCP, or an undetermined sink. CPF Oxon degradation was assumed to proceed to TCP
or an undetermined sink. TCP degradation was assumed to proceed to an undetermined
sink. Using this kinetic schema, the software then fits the data to the proposed
pathways, returning rate constants and relative weights of each pathway.
4.3.3 Results
4.3.3.1 TiO2 Characterization and Reactivity
Sonicating P25 in MD produced a narrow distribution of aggregates with an
average radius of 94.7 nm and a polydispersity index (PDI) of 0.06 as determined by
DLS, and confirmed with TEM. The resulting suspensions were stable, with no
123
aggregation observed over 6 h via time resolved DLS (Figure 57). TiO2 aggregates were
negatively charged at pH 7.5, with an electrophoretic mobility of -3.79 ± 0.11 x10-8 m2 V-1
s-1 (ζ potential of -48.4 ± 1.4 mV), and the presence of the media decreased the isoelectric
point from the point of zero charge of bare P25 (pH ~6.5) to a pH of approximately 1.5
(Data in Appendix C). Both �OH and O2�- generation was observed, with production
rapidly reaching steady state. The use of N-AC and SOD effectively quenched �OH and
O2�- production, respectively (Figure 58).
4.3.3.2 CPF Interactions with the TiO2 Surface
No significant loss of CPF from solution is observed under dark conditions for
either of the NP concentrations or the controls (Figure 28). Concentrations of CPF
observed in extracts were minimal, though no trend was observed either with increased
time or TiO2 concentration. Additionally, this occurs in the control samples as well as in
the presence of TiO2 and likely represents sorption of CPF to the glass walls of the
sampler vials after the aliquot has been taken. When combining the supernatant and
extract concentrations, CPF recovery compared to t = 0 for both dark and control
samples ranged from 89.5 to 101.5%.
124
0 500 1000 1500Time (min)
0
50
100
150
200
250
300
Chl
orpy
rifos
(ng/
ml)
0 mg L-1 Dark20 mg L-1 Dark40 mg L-1 Dark20 mg L-1 UV
Figure 28: Supernatant CPF concentration vs. time as a function of TiO2 concentration (0 - 40 mg L-1 dark and 20 mg L-1 UV).
4.3.3.3 CPF UV/TiO2 Degradation
Suspensions of CPF and TiO2 were exposed to UV light to initiate
photodegradation. Approximately 50% of the CPF was removed after 9 h given the
conditions of the photoreactor setup and TiO2 concentration (20 mg L-1). CPF can be
susceptible to base catalyzed hydrolysis, however no degradation products were
observed for any of the dark series (Data in Appendix C). Additionally, the working
stock solutions showed no decrease in concentration with time (Data in Appendix C).
125
Figure$2$
0 500 1000 15000
100
200
300
400
500
600
700
800
900
1000
Time (min)
nM
CPFCPF OxonTCPSum
t1/2 (min) CPF 443 CPF Oxon 6,150 TCP >10,000
Figure 29: CPF and degradation product concentration vs. time in suspensions of 20 mg L-1 TiO2 exposed to UV.
Degradation of CPF and CPF Oxon and TCP production are shown in Figure 29.
Summing the three analytes returns an acceptable mole balance (90.8-105.3% recovery),
indicating that the formation of additional primary degradation products is unlikely.
After 24 h of continual UV exposure, 81% of CPF was degraded, and the rate of CPF
Oxon formation slowed, though a decrease in CPF Oxon concentration was not observed
over the sampling period.
Data were fit using the CAKE software package and followed first order reaction
kinetics (Data in Apendix C). The half-life of CPF was calculated to be 443 minutes
under the given conditions, while half-lives for CPF Oxon and TCP were calculated to be
on the order of hundreds of hours (Figure 29, inset). CPF Oxon is the predominant
126
Figure3
!
UV/TiO2!oxidation! N
OP
O
OO
CH3
CH3
Cl
Cl
Cl
N
OH
Cl
Cl
Cl
OHCH3OH
PO
OO
CH3
CH3
OHP
OH
O
OH
hydrolysis!
chlorpyrifos!oxon!
t1/2!~!1!d!(TiO2(s)!surface)!
N
OP
O
SO
CH3
CH3
Cl
Cl
Cl
chlorpyrifos!
Figure'2.!Initial!products!expected!after!degradation!of!chlorpyrifos!by!UV/TiO2!application!3,5,6-Trichloro-2-pyridinol
!
UV/TiO2!oxidation! N
OP
O
OO
CH3
CH3
Cl
Cl
Cl
N
OH
Cl
Cl
Cl
OHCH3OH
PO
OO
CH3
CH3
OHP
OH
O
OH
hydrolysis!
chlorpyrifos!oxon!
t1/2!~!1!d!(TiO2(s)!surface)!
N
OP
O
SO
CH3
CH3
Cl
Cl
Cl
chlorpyrifos!
Figure'2.!Initial!products!expected!after!degradation!of!chlorpyrifos!by!UV/TiO2!application!
Chlorpyrifos
!
UV/TiO2!oxidation! N
OP
O
OO
CH3
CH3
Cl
Cl
Cl
N
OH
Cl
Cl
Cl
OHCH3OH
PO
OO
CH3
CH3
OHP
OH
O
OH
hydrolysis!
chlorpyrifos!oxon!
t1/2!~!1!d!(TiO2(s)!surface)!
N
OP
O
SO
CH3
CH3
Cl
Cl
Cl
chlorpyrifos!
Figure'2.!Initial!products!expected!after!degradation!of!chlorpyrifos!by!UV/TiO2!application!
Chlorpyrifos Oxon
Unidentified Products
Unidentified Products
Unidentified Products
Figure 30: Potential degradation pathways considered by CAKE software.
degradation product produced via oxidation of the sulfur in the phosphorothioate
group, and this pathway represents 84% of the calculated CPF Degradation. TCP
formation, formed through ester cleavage, makes up the balance at 16%. In agreement
with the mole balance achieved, no significant loss of CPF to an unidentified sink was
identified, though this pathway was included as an option as illustrated in Figure 30.
The further oxidation of CPF Oxon to TCP was not quantifiable, most likely due to
reactor conditions resulting in little CPF Oxon degradation over 24 h.
4.3.3.4 Impact of UV/TiO2 on Bacteria
The hydroxyl radical is primarily responsible for bacterial inactivation as
determined through the use of ROS quenchers (Data in Appendix C). The addition of
SOD did not alter the inactivation rate of either A. baumannii or B. subtilis, indicating
127
superoxide was not generated at sufficient concentrations to contribute to bacteria
toxicity. Conversely, no inactivation was observed when bacteria were irradiated in a
TiO2 suspension to which the �OH quencher N-AC was added, further implicating �OH
as the primary active agent. Strong inactivation was observed for both A. baumannii and
B. subtilis when exposed to UV irradiated TiO2. For A. baumannii, fewer than 10% of CFU
remained after 60 minutes of irradiation, while B. subtilis was completely inactivated
over the same timeframe.
Figure 4
0 10 20 30 40 50 600
0.2
0.4
0.6
0.8
1
1.2
1.4B.subtilis Inactivation
Time (min)
N/N
0
UV+ SOD N−AC D+ UV−
0 10 20 30 40 50 600
0.2
0.4
0.6
0.8
1
1.2
1.4A.baumannii Inactivation
Time (min)
N/N
0
UV+SODN−ACD+UV−
a) b)$
Figure 31: Inactivation of bacteria (A. baumannii and B. subtilis) under the following conditions: (UV+) 20 mg L-1 exposed to UV, (SOD) 20 mg L-1 TiO2 exposed to
UV in the presence of SOD, (N-AC) 20 mg L-1 TiO2 exposed to UV in the presence of N-AC, (D+) 20 mg L-1 TiO2 in the dark, and (UV-) 0 mg L-1 TiO2 exposed to UV.
TEM of bacteria exposed to UV light both in the presence and absence of NPs for
60 minutes show small clusters of nanoparticles observed in close proximity to ruptured
segments of cell membranes (Figure 32). These images further support the evidence
128
indicating that locally generated reactive oxygen species are most likely responsible for
loss of membrane integrity and cell death.
Figure 32: TEM images of B. subtilis with 20 mg L-1 TiO2 (a) prior to UV exposure and (b) after 60 minutes UV exposure. Scale bars are 500nm.
To investigate if the sorption of an organic contaminant affects bacterial
inactivation (either positively or negatively), bacteria were irradiated in the UV/TiO2
system in the presence of CPF at 375 µg L-1. Inactivation was measured after allowing
CPF to sorb to the mineral surface for 60 minutes in the dark and compared to
inactivation from TiO2 not exposed to CPF. The presence of CPF did not produce a
significant difference in inactivation from suspensions without CPF, as shown in
Appendix C, Figure 61.
129
4.3.3.5 Impact of CPF and Degradation Products on Bacteria
The presence of CPF had no appreciable impact on bacterial inactivation rates
(Data in Appendix C), either when irradiated immediately or following 60 minutes of
mixing in the dark. Additionally, no inactivation was observed following incubation of
bacteria for up to 1 h in the resulting combination of CPF, TiO2 nanoparticles, and CPF
degradation products created by 60 minutes of UV irradiation. Furthermore, CPF by
itself did not inactivate A. baumannii or B. subtilis over 60 minutes of exposure. Finally,
CPF oxon, and TCP, the primary degradation products, were dosed at the equivalent
molar concentrations used for CPF to simulate complete transformation of the pesticide
to a degradation product. Neither CPF Oxon nor TCP was found to be acutely toxic to
either A. baumannii or B. subtilis (Data in Appendix C).
4.3.3.6 Impact of Bacteria and CPF on TiO2 Reactivity
To determine the impact that the presence of additional constituents in
suspension had on reactivity, �OH production was measured in both the presence and
absence of CPF and B. subtilis. CPF was added to a TiO2 suspension and mixed briefly in
the dark prior to the addition of TA and exposure to UV-A for 5 minutes. The resulting
suspensions contained 1.098 µM CPF and 125 µm TA. 2-HTA fluorescence represents
the amount of hydroxyl radicals that are detected in the bulk. As can be seen in Figure
33(a), the presence of CPF decreased the amount of hydroxyl radicals detected by TA by
88±1.3%. No change in reactivity was observed compared to initial values when CPF and
130
TiO2 were mixed up to 60 minutes before the addition of TA and illumination (Figure
33(b)). This is consistent with the lack of sorption observed in Section 5.3.2. Furthermore,
time resolved DLS did not indicate TiO2 aggregation in the presence of CPF (Data in
Appendix C).
Figure$6$
0min 30min 60min0
0.2
0.4
0.6
0.8
1
Rel
ativ
e Fl
uore
scen
ce (F
SU)
Hydroxyl RadicalSuperoxide
Dark UV subtilis CPF CPF+subtilis0
0.2
0.4
0.6
0.8
1
Rel
ativ
e Fl
uore
scen
ce (F
SU)
a) b)
Figure 33: Reactivity measurements of 20 mg L-1 TiO2 exposed to UV. (a) Detected �OH in the presence of various constituents. (b) Detected O2�- and �OH in
the presence of CPF over 60 minutes.
Similarly, B. subtilis was added to a suspension of TiO2 at a concentration of 108
cfu ml-1 and reactivity was subsequently measured via TA. The presence of bacteria
decreased 2-HTA fluorescence by 56±10%. When both CPF and B. subtilis are present in
suspension, fluorescence decreased by 94±0.4%. This value is greater than dark controls
(96±0.8%), indicating that a small amount of hydroxyl radicals are still detected by the
probe compound.
4.3.3.7 Impact of Bacteria on CPF Degradation
CPF degradation tests were performed in the presence of B. subtilis at 20 mg L-1
131
TiO2. Given the reactivity measurements observed in Section 5.3.6, it was expected that
TiO2 association with the bacteria would reduce the degradation rate of CPF. In fact, no
difference was seen in CPF loss rates with bacteria compared to without (Figure 34).
Dark controls indicated negligible CPF Oxon or TCP production, and 92-115% recovery
over 24 h. (Data in Appendix C).
Figure$7$
0 500 1000 15000
0.2
0.4
0.6
0.8
1
Time (min)
Chl
orpy
rifos
(C/C
0)
TiO2 onlyTiO2 + B.subtilis
Figure 34: Degradation of CPF in the presence and absence of bacteria.
4.3.4 Discussion
4.3.4.1 TiO2 Photocatalysis
Photodegradation rates depend upon many factors (e.g. UV flux, contaminant
concentration, and photocatalyst loading) making direct comparisons of studies
difficult. In this work, reactor conditions were selected to better observe the interactions
between TiO2, CPF, and bacteria. The CPF concentration used in this study (375 µg L-1)
132
was selected due to the low aqueous solubility of the pesticide. However there is great
variability reported in the literature for CPF solubility, Solubility values as low as 300 µg
L-1 have been reported, although Gebremariam et al. report an average of 600±330 µg L-1
in distilled water [321]. Frequently, degradation studies have far exceeded this
solubility, with concentrations ranging from mg L-1 to hundreds of mg L-1, often by using
significant concentrations of co-solvents [346, 347, 367, 368]. By selecting a lower
concentration of the pesticide, we minimize unexplained losses arising from the
compound’s hydrophobicity and the use of co-solvents that may quench produced ROS.
In addition to large CPF concentrations, degradation studies have often
employed high catalyst loading with the goal of maximizing the photodegradation rate
[344]. Generally, ROS production is seen to improve with increasing TiO2 concentration
up to the point when light penetration becomes the limiting factor in photoexcitation.
The concentration at which this optimal production occurs varies widely depending on
experimental setup [347, 369]. Even at the lower end, nanoparticle particle
concentrations frequently employed are an order of magnitude greater than the 20 mg L-
1 applied in the current study. Using a lower catalyst loading effectively prolongs the
degradation process, allowing better insight into the potential pathways of CPF removal
from solution.
133
4.3.4.2 CPF Adsorption
While high affinity of CPF for soil particulate matter has often been observed
[322, 348, 370], generally, few attempts have been made to identify sorption versus
degradation of CPF using TiO2. Devi et al. report an 8% loss of CPF from suspension
after mixing in the presence of 150 mg anatase for 10 minutes in the dark, which is
ascribed to adsorption [332]. It is unclear, however what concentrations of CPF were
used in this study, and an attempt to extract the pesticide off of the TiO2 to validate the
proposed does not appear to have been made.
Studies focusing on the interactions and surface catalyzed hydrolysis of
chlorpyrifos methyl and other related pesticides using metal oxides have generally
found little to no sorption. Nair et al., looking at adsorption of CPF onto Al2O3 found
complete removal of the pesticide when the mineral surface was coated with Au and Ag
NPs, but that no significant adsorption occurred with bare Al2O3 spheres [348].
Furthermore, Torrents et al. observed the slow loss of chlorpyrifos methyl from solution
in a 10 g L-1 TiO2 suspension, but concluded this decrease was likely due to hydrolysis
rather than sorption [349]. The rate of hydrolysis by TiO2 was greater than that for
equivalent concentrations of Al2O3 or FeOOH. Clausen and Fabricius, utilizing α- and γ-
FeOOH, showed that absorption, while significant for ionic pesticides, was negligible for
nonionic pesticides [350].
134
These studies support our current observations for the nonionic chlorpyrifos,
and suggest that the use of TiO2 NPs alone will not contribute practically to CPF
adsorptive removal from solution in a remediation scenario. As a result, loss of CPF is
likely to occur almost entirely through photocatalysis. The timeline for TiO2 surface
catalyzed hydrolysis presented by Torrents et al. indicate that the order of days is
necessary for tens of g L-1 of NPs to degrade CPF appreciably, while this study, using a
thousand-fold less TiO2, is capable of achieving a 50% reduction in CPF in 9 h [349].
Furthermore, the lack of specific interactions implies that other components of natural
waters, such as bacteria or NOM, which may show a stronger affinity for TiO2, are likely
to further limit access to the TiO2 surface.
4.3.4.3 CPF Degradation and Oxidation Products
Duirk et al. observed both CPF Oxon and TCP formation from oxidation of CPF
using HOCl as the oxidant [371]. The relative amounts of the degradation products
produced are in good agreement with the current study, in which the fraction of CPF
transformed via oxidation of the sulfur, far exceeds cleavage of the ester bond to create
TCP. Devi et al. present a similar pathway [332]. Thus, while ester cleavage may be the
predominant pathway for surface catalyzed hydrolysis, oxidation produces an
intermediary, which is then capable of forming TCP or further degrading via an
alternative pathway.
135
4.3.4.4 Comparison of TiO2 Reactivity in the Presence and Absence of CPF
The formation of ROS occurs at the surface of the TiO2 nanoparticle, which then
diffuses away from the surface into the bulk solution, nonspecifically reacting with
targets it comes across [40]. As each compound in solution competes for a fraction of the
finite amount of ROS generated, delivery depends not only on individual reaction rates
but also proximity to the NP surface. Because �OH generated by heterogeneous
photocatalysis is produced at a specific point (versus homogenous photocatalysis, e.g.,
hydrogen peroxide which is uniformly mixed in aqueous solutions) the closer a target is
to the photocatalyst, the greater the subsequent radical dose. The use of a probe
compound such as TA is capable of providing a benchmark of the �OH concentration in
the bulk solution. In this study, CPF is observed to outcompete TA for hydroxyl radicals,
as evidenced by the strong decrease in fluorescence observed in Figure 33 when CPF is
introduced. This could result either from a greater reaction rate with �OH or from closer
association with the TiO2 surface.
The �OH reacts with TA and CPF at roughly equal rate of 3.3x109 M-1 s-1 and
4.4x109 M-1 s-1, respectively [88, 372]. Given that the concentration of CPF present in
solution (1.098 µM) was two orders of magnitude less than that of TA (125 µM), one
would expect the presence of CPF to have a negligible impact in 2-HTA formation, and
thus fluorescence. That this is not the case implies a hierarchy of affinity for the NP
surface that brings CPF closer to the radical generator and increases the probably of CPF
136
oxidation over TA.
Though sorption of CPF onto the surface of TiO2 is not observed (Section 5.3.2),
relatively weaker hydrophobic interactions are likely to draw the nonionic CPF towards
the NP surface. Additionally, the two partially positive ethyl functional groups attached
to the more electronegative phosphorothioate in CPF will produce a small dipole
moment that may encourage interaction with the negatively charged NP surface. These
ion-dipole interactions are likely to be smaller in magnitude compared to ion-ion
repulsive forces between TiO2 and the negatively charged carboxylic groups on TA (pKa
3.5, 4.5) [88], but the effect will be the same in either case. These electrostatic interactions,
coupled with the hydrophobic interactions of CPF, will draw CPF molecules closer to
the surface of the negatively charged TiO2 and relegate TA to a position further out in
the bulk solution resulting in a greater exposure to hydroxyl radicals for the pesticide
and effectively shielding the water-soluble probe compound.
4.3.4.5 Reactivity in the Presence of CPF and Bacteria
The nature of these interactions drawing CPF to the NP surface, however, is not
so strong that it displaces bacteria or alters the rate of inactivation, as observed in
Appendix C, Figure 61. These data, coupled with TEM images in which NPs are
observed at the surface of bacteria both prior to and following UV exposure, indicate
that NPs attach to the cell providing a direct pathway for �OH delivery. These results
imply that the CPF remediation process itself (i.e. photoactivated TiO2) will likely have a
137
greater impact on microbial viability than either the pesticide or its degradation
products at concentrations below the solubility limit given the lack of acute toxicity
observed.
Conversely, nor does the presence of bacteria impact CPF degradation over 24 h.
This implies that, despite TiO2 affinity for the bacterial surface, there remains sufficient
�OH production for CPF degradation to continue unaffected. These data are in contrast
to the decrease in detected �OH by TA in Section 4.3.3.6, and further indicate that
differences in relative affinity for the NP act as significant deciders of delivered ROS
dose.
4.3.5 Conclusions
The picture that emerges from the given system is one in which components of
the water environment are impacted relative to their affinity for the TiO2 NPs. Thus,
association of TiO2 with bacteria and with the hydrophobic CPF results in a greater �OH
dose delivery than for the probe molecule TA for which there is less affinity. Arising
from this are the implications that while investigating reactivity in simplified systems
can provide a benchmark of NP reactivity, it may also overestimate photocatalytic
efficiency in realistic environments, especially for hydrophilic contaminants, and that
more research in complex systems should be undertaken to better understand the role
that non-targeted entities may play in remediation scenarios.
138
5. Conclusions The experiments described in this dissertation were designed to improve our
understanding of how changes in NP surface chemistry impact the photoreactivity of a
nanoparticle. This has been carried out through investigations focusing on surface
functionalization, specific and non-specific interactions with anions in solution, and
association with biological surfaces and organic molecules. NPs were characterized in
terms of aggregate morphology (size and structure) and particle stability in addition to
photoreactivity endpoints (radical probes, microbiological impact, and contaminant
degradation). The results of this work have highlighted the important role that the NP
surface will play in determining reactivity in a natural aquatic environment and have
supported the hypotheses outlined in Chapter 1, namely 1) that functionalization of
FNPs will reduce aggregation and impact ROS generation, 2) that water chemistry and
interactions with organic or ionic compounds at the NP surface will determine the
reactivity of the system, and 3) that multiple, coexisting constituents will compete for
ROS delivery.
5.1 Bacteriophage Inactivation by UV-A Illuminated Fullerenes: Role of Nanoparticle-Virus Association and Biological Targets
Engineered functionalization of FNPs was shown to impact ROS production,
reduce aggregation, and alter viral proximity, resulting in the following hierarchy of
MS2 virus inactivation: aqu-nC60 < C60(OH)6 ≈ C60(OH)24 < C60(NH2)6, which
139
followed Chick-Watson inactivation kinetics. Changes to FNP surface functionalization
were observed to significantly impact photoreactivity, with the addition of 24 hydroxyl
groups to the C60 cage influencing the surface charge and increasing singlet oxygen
production. As a result these fullerol aggregates were smaller and more stable than nC60,
and are expected to have a longer lifetime in the water column, which, combined with
the increase in observed singlet oxygen production, would result in greater potential
exposure and hazard, respectively.
The addition of amine functional groups both decreased the size of aggregates
and the magnitude of surface charge while increasing the amount of ROS produced,
suggesting that antiviral properties of fullerenes can be increased by adjusting the type
of surface functionalization and extent of cage derivatization, thereby increasing the
ROS generation rate and facilitating closer association with biological targets.
When comparing the extent, rather than type, of derivatization, few differences
in aggregate characteristics were observed between C60(OH)6 and fullerol, that latter of
which has often been used in research due to its relative ease of suspension. While the
average fullerol aggregate was smaller than C60(OH)6, measurements of electrophoretic
mobility are indistinguishable at the circumneutral pH of PBS. Furthermore, singlet
oxygen generation of C60(OH)6 was only slightly less than that of fullerol. Thus, the
hydroxylation of fullerenes expected to occur to in natural waters following UV
140
exposure will rapidly result in products that behave similarly to their more
functionalized counterparts, capable of generating significant amounts of ROS.
What remains unclear is how engineered functional groups will impact the
phototransformations that are expected to occur. As FNPs are tailored to specific
products or uses, highly functionalized FNPs will become increasingly more prevalent,
with pristine fullerenes becoming increasingly less likely to be released into the
environment. This work implies, however, that these advanced FNPs will retain a
significant degree of photoreactivity despite any transformations (e.g. degradation,
oxidation, cleavage) that occur to large functional groups existing on the surface. How
the aggregation state and affinity of a specially functionalized FNP for a biological
surface changes with these transformations may be the largest factors in determining the
delivered dose in natural waters.
5.2 Influence of Inorganic Anions on the Reactivity of TiO2 Nanoparticles
Unlike the functionalization of fullerenes, changes to the crystalline lattice of
TiO2 nanoparticles following environmental release are unlikely, though this may
indeed occur with some metal oxide nanoparticles, such as nano ceria, due to redox
chemistry. While intrinsic properties of TiO2 are expected to remain constant, surface
interactions with organic matter, colloids, or, as shown in this work, inorganic anions
present in water will significantly impact the reactivity of the system.
141
The influence of anions on surface reactivity and bulk ROS production has been
demonstrated though the response is not always straightforward, as the greater impact
of phosphate on direct hole oxidation compared to hydroxyl radical generation
demonstrates. Thus, determining the potential hazard of a NP depends at least partially
on the mode of action. For hydrophobic entities sorbed to the NP surface, a decrease in
bulk ROS levels due to radical quenching by an anion may have no impact in perceived
reactivity.
Furthermore, specific interactions with anions such as phosphate and carbonate
undergoing in inner sphere ligand exchange have been shown to stabilize NPs due to
increased negative surface charge. EPM was measured versus pH to demonstrate the
specific interactions of phosphate and carbonate for which the pH of the IEP was
concentration dependent. No change in IEP was observed in the presence of nitrate or
chloride illustrating the lack of specific interactions. As a result, small concentrations of
these anions may increase the overall risk of a nanoparticle, evidenced by the increase in
both ROS generation and particle stability in waters containing 0.5 mM phosphate and
carbonate compared to nitrate or chloride. Ultimately, this work demonstrates that to
understand the reactivity of TiO2 nanoparticles, the aquatic environment must be
considered. Furthermore, this research suggests that reactivity is not static; that the
transition of NPs from water of a given ionic composition to another of similar ionic
strength may result in either an increase or decrease in reactivity.
142
5.3 Chlorpyrifos Degradation via Photoreactive TiO2 Nanoparticles: Assessing the Impact of a Multi-Component Degradation Scenario
Chapter 4.3 studied the influence of ancillary components on bacterial
inactivation or contaminant degradation by TiO2, and ROS delivery was shown to
depend on the coexisting constituents of a suspension. The presence of either bacteria or
CPF strongly decreased bulk measurements of hydroxyl radicals, with almost no �OH
detection in solutions containing both, indicating that proximity to the NP surface is
important in determining ROS dose. However, the presence of bacteria had no impact
on CPF degradation rates, nor did the presence of CPF diminish the rate of bacterial
inactivation. Thus, while TEM images show aggregates of TiO2 sorbed to the surface of
bacteria, sufficient hydroxyl radicals are generated and escape delivery to the bacteria,
reacting instead with CPF.
Ultimately, the implication of this study is that components in suspension will be
impacted by photoreactive nanoparticles relative to their affinity for the NP surface. This
agrees well with findings described in Chapter 4.1 in which various FNPs were observed
to have differing affinities for a viral surface. This point source nature of photoreactive
NPs, in which ROS is not uniform in solution, is a significant complicating factor in
assessing NP risk and emphasizes the importance of the physical characteristics of an
aggregate. This work suggests that estimating risk based off of simplified systems may
poorly estimate photocatalytic efficiency in realistic environments; for hydrophilic
143
targets, the presence of constituents with greater affinity for the NP surface is likely to
diminish the delivered dose.
5.4 Summary of Conclusions
This dissertation has demonstrated that the reactivity of a system is largely
determined by the surface chemistry of a NP. These surface interactions may be
intentional, such as functionalized FNPs, or incidental and resulting from the aquatic
environment into which a NP is introduced. Additionally, the photocatalytic impact of a
NP will depend on the characteristics of potential targets as on the characteristics of the
NP itself.
5.5 Continuations of This Work
The studies described herein have necessarily employed simplified systems in
order to understand the trends that impact reactivity. By no means have all the
components of natural waters that may influence reactivity been considered in this
work; prime candidates for future investigation include NOM and colloids such as clays
or silicates, as well as engineered waters with greater complexity.
5.5.1 Natural Organic Matter
NOM presents a particularly intriguing case, as it can act both as a sensitizer and
quencher of radicals. Studies have shown NOM exposed to sunlight produces sufficient
amounts of ROS to inactivate bacteria and degrade organic compounds. Additionally,
NPs coated in NOM have shown increased stability and mobility in column studies.
144
Understanding the conditions in which NOM may increase or decrease the total ROS of
a NP containing system will greatly contribute to predictions of risk related to NP
release. Furthermore, comparing the relative contribution of NP and NOM on ROS
production can provide a benchmark of environmental impact.
5.5.2 Natural Colloids
Colloids, ubiquitous in rivers lakes and streams, are anticipated to outnumber
NPs in all but the most concentrated release scenarios, and it is reasonable to expect
heteroäggregation to dominate in most cases. Currently, heteroäggregation studies have
been limited to the impact on aggregate formation and morphology. The influence of
aggregate size and morphology in homoäggregates on ROS production (e.g., shielding
of UV light and radical quenching) has been demonstrated, but the formation of
heteroaäggregates is likely to add further pathways of excitation (e.g., charge transfer,
electron or hole trapping).
5.5.3 Photoreactivity in Natural Systems – Towards Functional Assays
While investigations of simplified systems are important tools in understanding
photoreactivity, the results presented in Chapter 4.3 illustrate the complexity that arises
in conditions containing even mild resemblance to natural waters. Developing a
comprehensive understanding of reactivity will require testing FAs in agreed upon
representative environmental conditions. Among the necessary tasks then, are to
determine reasonably simple reactivity endpoints and characterize anticipated exposure
145
conditions and relevant forms of released NPs. Water chemistry will impact NP
aggregation, dissolution, fate and transport, and these physicochemical changes to NPs
should continued to be studied to better inform risk assessment. Organismal impacts
may vary between levels of complexity (e.g., fish versus bacteria) or even at the same
level (e.g., gram positive versus gram negative bacteria), highlighting the need to
identify appropriate endpoints.
5.5.3.1 Functional Assay Endpoint Considerations
Given the complexity in assessing potential toxicity, an appropriate FA endpoint
such as exogenous ROS production could help identify combinations of NPs and
situations for which hazard may be high. The goal should not be to measure each
specific impact of a NP, but to identify and investigate the most likely impacts or types
of NPs associated with higher risk. Similarly, instead of testing each susceptible
organism under a number of conditions, a stronger and more productive approach
would be to look at photoreactivity in cases for which that’s the likely mode of action.
An example would be the release into aqueous environments of photocatalytic
metal oxides that have the potential to be toxic to a number of microorganisms. Risk
assessments of these particles ought reasonably to take into account their fate and
transport in the anticipated environment as well as endpoints related to photoreactivity.
The potential impact of these particles is unlikely to be the same if one considers their
introduction into a stream or lake versus a WWTP, in which the majority of NPs have
146
been found to partition into the biosolids and may be later land applied as part of soil
amendments. For these scenarios, the appropriate FAs would differ.
TiO2 and Fullerenes are attractive model NPs for establishing FAs, as they are
generally stable and mostly non-toxic in the absence of light. However, as has been
demonstrated, their photoreactivity results in toxicity, measurements of which could
serve as a realistic FA. In this dissertation, investigations into ROS generation have given
a better understanding of potential for toxicity and degradation, measurements that
incorporate the significant influences of aggregate size and morphology. Measuring ROS
production does not invalidate the need for NP characterization or system property
measurements; but with the understanding that NP aggregation will impact the
underlying phenomena, it may be sufficient to measure ROS production as a
determinant of hazard for a given set of conditions. Using this as a quantifiable
endpoint, we can easily and systematically investigate what affects reactivity of NPs in
the environment.
5.5.3.2 Environmental Release Considerations
Challenges to predicting risk for photoreactive nanoparticles include
uncertainties in both aquatic chemistry and light exposure. While the intensity of
sunlight can be largely predicted on a coarse (daily) scale based on season, geographic
location, etc., results presented in this dissertation highlight the fact that NP
transformations such as aggregation resulting from the aquatic chemistry are likely to
147
take place on a much finer scale (minutes to hours). As a result, NPs released into
exposed waters at noon on a sunny day will present a greater hazard than those released
into shaded waters or late in the evening. It may be sufficient to study reactivity under
the worst-case scenario (i.e., using the greatest reasonable light flux) and apply a “light
factor” ranging from 0 to 1 to adapt experimental results to anticipated release
conditions. However, while this approach appears to be straightforward for unchanging
NPs or aggregates, the impact of transformations my not be easily separated from UV
exposure. It would be useful to measure reactive endpoints at high, low, and average
light exposure conditions to determine what can be reasonably estimated.
Waters containing stabilizing agents such as phosphate, carbonate or natural
organic matter are likely to increase the duration of NPs in the water column, thereby
increasing the net flux of light to which they are exposed. Furthermore, water chemistry
may impact the type of ROS produced. For example, irradiated FNPs in the presence of
an electron donor favor superoxide production over single oxygen. In high nitrate
waters, peroxynitrate may be generated or the carbonate radical in high carbonate
waters. These secondary radicals, which have less oxidizing capacity, are largely
dependent on a priori �OH generation and are less likely to necessitate quantification,
though in waters where they are the predominant form of ROS to which organisms are
exposed, they could be appropriate endpoints.
148
The identification of realistically anticipated environmental conditions would
greatly facilitate the design of photoreactivity FAs. Initially, multiple ROS endpoints
should be tested to adequately describe the reactivity of a NP. Priority should be given
to the primary expected ROS endpoints (singlet oxygen for FNPs, superoxide and
hydroxyl radical for semiconductors) as well as measurements of surface reactivity
(direct hole oxidation). Ultimately the precise ROS measurement for a photoreactivity
FA will depend on the water chemistry of the release scenario.
5.5.3.3 Nanoinformatics, Data Visualization, and Comparison
This work has shown that changing the anion concentration significantly impacts
ROS generation in simplified systems. As we consider more realistic conditions,
however, it will become important to rank or compare reactivity between greatly
differing water chemistries. For example, it will be more difficult to understand and
predict the reactivity of waters in which ionic strength, suspended solid content, and
organic matter all vary.
The further development and employment of nanoinformatics databases, such as
those being developed within CEINT (NIKC, N4mics) or by the EPA (NaKnowBase) will
provide platforms for data visualization and comparison that enable complex
relationships to be discerned from multiple studies consisting of multiple parameters.
Comparisons of NP response between these relevant environmental conditions will help
to identify the primary variables in determining reactivity. This, coupled with the
149
establishment of reasonably simple, appropriate FAs will move the process of NP risk
assessment away from a case-by-case handling of each NP and enable regulation to keep
pace with the advancements of nanotechnology.
150
Appendix A – Supporting Information for Chapter 4.1: Bacteriophage Inactivation by UV-A Illuminated Fullerenes: Role of Nanoparticle-Virus Association and Biological Targets Suspension Preparation
A colloidal suspension of fullerene (aqu-nC60) in ultrapure water (UPW) was
prepared by sonicating 200 mL of UPW (resistivity > 18 MΩ cm and TOC< 30 µg/L)
containing 100 mg C60 powder using a high-energy probe (S-4000, Misonix, Qsonica,
LLC, Newtown, CT) for 10 hours (pulse time: ON for 10 min; OFF for 20 min) in the
dark. Following the dispersion, the suspension was filtered through a 100 nm nylon
membrane filter (Magna, GE Osmonics, Minnetonka, MN). The filtrate was then
concentrated via rotavaporator (Büchi, Flawil, Switzerland). DLS measurements of the
suspension pre- and post-concentration indicated no change in average aggregate size.
The number weighted and unweighted particles size distributions of FNPs in aqueous
suspension are shown in Figure 35.
151
!
0.1 1 10 100 10000.0
0.2
0.4
0.6
0.8
1.0
(a)
C60
(NH2)6
0.96±0.23 nm
C60(OH)242.23±0.19 nm
C60(OH)64.23±0.11 nm
aqu-nC6040.25±0.23 nm
Num
ber w
eigh
ted
inte
nsity
Nanoparticle radius (nm)
!
0.1 1 10 100 1000 100000.0
0.2
0.4
0.6
0.8
1.0
(b)
C60(NH2)60.13±0.34 nm (6.27%)54.39±1.84 nm (93.73%)Avg.=31.19±2.31 nm
C60(OH)243.33±0.29 nm (5%)80.55±0.57 nm (95%)Avg.=68.67±0.89 nm
C60(OH)64.58±0.12 nm (7.34%)43.17±0.13 nm (92.66%)Avg.=36.62±0.60 nm
aqu-nC6085.66±0.52 nm (68.67%)3168±0.59 nm (31.33%)Avg.=265.5±1.76 nm
Unw
eigh
ted
inte
nsity
Nanoparticle radius (nm)
Figure 35: (a) Number weighted intensity and (b) unweighted intensity of fullerenes in suspension measured using dynamic light scattering at 90 degrees
scattering angle.
Electrophortic Mobility Measurements
The Electrophoretic mobilities of the FNP suspensions as well as of MS2 virus are
summarized in Table 6. As can be seen, all suspensions were negatively charged under
our experimental conditions.
Table 6: Electrophoretic mobility and zeta potential of FNP suspensions and MS2 virus.
EPM (m2/V.s) x10-8 ZP (mV) Suspension avg std (n=6) avg std (n=6) C60(NH2)6 -1.09 0.10 -13.9 1.2 C60(OH)6 -3.65 0.30 -46.6 3.9 aqu-nC60 -1.76 0.29 -22.4 3.7 C60(OH)24 -3.57 0.30 -45.6 3.8 MS2 -1.09 0.09 -13.9 1.1
Standard Curve Preparation
152
Singlet oxygen production was determined by comparing the measured
fluorescence of SOSG in fullerene suspensions versus a Rose Bengal (RB) standard
curve. All SOSG measurements were performed in the absence of MS2 virus. RB
standards produced a linear fluorescence response over a concentration range of 0.005 –
0.1 µM shown in Figure 36.
!
0.00 0.02 0.04 0.06 0.08 0.100
100
200
300
400
500
y= 3511.5x + 93.8
ΔFS
U/m
in
Rose bengal (µM)
Figure 36: Change in fluorescence measured over one minute irradiance versus rose bengal standard concentration. Error bars indicate the standard deviation of
triplicate measurements.
The rate of photons hitting the sample in photons per second was calculated as
follows:
rphoton =irr ⋅a ⋅λh ⋅c
(Eq 34.)
where irr is the irradiance in W/cm2, a is the illuminated area of the reaction vial, c is the
speed of light, λ is the wavelength, and h is plank’s constant. Combining this with the
RB standard solution absorption at 365nm, ABS365, the quantum yield, Φ, and
153
Avagadro’s number, NA, gives the rate of singlet oxygen production, r1O2, in molar
concentration per time:
r1O2 =rphoton (1−10
−ABS365 )φV ⋅NA
(Eq 35.)
Figure 3 shows the fluorescence response of SOSG versus singlet oxygen
production in the RB standards. Concentrations of irradiated fullerene suspensions were
adjusted to maintain fluorescent responses within this linear standard range
!
0.0 0.2 0.4 0.6 0.8 1.00
100
200
300
400
500
600
y=(404.33 ± 28.87)x + (93.80 ± 14.10)
R2 =0.99
ΔFS
U/m
in
1O2 generation rate (µM/min)
Figure 37: Rate of change in fluorescence measured over one-minute irradiance vs. calculated singlet oxygen production by the RB standards indicated in Figure 39 and calculated by Equations 34 and 35. Error bars indicate the standard deviation of
triplicate measurements.
The UV-A Irradiance of the RB standards was 2.3 mW/cm2 as measured by a
UVX radiometer (UVP, Inc. Upland, CA). The output spectrum of the UV bulbs used in
the study are shown in Figure 38.
154
Figure S4. Excitation spectra of the UV bulb used in experiment (provided by Koninklijke Philips Electronics N.V.)
Wavelength*(nm)*
amount*of*energy*(watts/cm^2)*
(joules/cm^2/sec)*
340$ 0.0000021$341$ 3.57509E,06$342$ 4.61162E,06$343$ 5.8859E,06$344$ 7.43303E,06$345$ 9.28777E,06$346$ 1.14829E,05$347$ 1.40469E,05$348$ 1.70023E,05$349$ 2.03622E,05$350$ 2.41288E,05$351$ 2.82904E,05$352$ 3.28198E,05$353$ 3.76727E,05$354$ 4.27868E,05$355$ 4.80824E,05$356$ 5.34632E,05$357$ 5.8819E,05$358$ 6.40285E,05$359$ 6.89639E,05$360$ 7.34959E,05$361$ 7.74993E,05$362$ 8.08585E,05$363$ 8.34731E,05$364$ 8.52629E,05$
Product data
Bulb T26Main Application Blacklight BlueUseful Life 5000 hrLife to 50% failures EM 8000 hrRated Lamp Wattage 15WTechnical Lamp Power 15.9 WLamp Voltage 54 VLamp Current 0.335 AMercury (Hg) Content 5.0 mgColour Code 108 [08 lead free glass]Colour Designation Blacklight BlueUV-A Radiation 100hr (IEC) 3.0 WUV-B/UV-A (IEC) 0.2 %
TL-D/08/108 G13
Cap-Base G13
TL-D/08/108
TL-D/08/108
2
25/2/2011
Figure 38: Output spectrum of the Philips 15W TL-D BLB bulbs used in the study. Output is centered at 365nm, and UV-A irradiance was measured at 2.3mW/cm2.
Inactivation Kinetics and Singlet Oxygen Dose
To determine the steady state singlet oxygen concentration, [1O2]ss, first the
accumulation of singlet oxygen in suspension can be modeled as follows
d[1O2 ]dt
= r1O2 − kd[1O2 ] (Eq 36.)
where kd is the rate of quenching in water (kd = 2.4x105 sec-1)[373]. Making the
steady state approximation, an equation for [1O2]ss, in moles, can be calculated as
[1O2 ]ss =r1O2kd (Eq 37.)
Thus, from the Rose Bengal standard suspensions, the fluorescence response due
to a given steady state singlet oxygen concentrations can be calculated as
dFSUdt
= k1O2 [1O2]ss[SOSG] (Eq 38.)
155
where FSU is the detected fluorescence resulting from the cumulative reaction of SOSG
with the produced 1O2 over the irradiance time t, k1O2 is the second order rate constant,
and [1O2]ss is the steady state 1O2 concentration. By using a sufficiently large [SOSG]
compared to produced 1O2 and short irradiance times, the concentration of unreacted
SOSG in solution can be assumed constant, giving the pseudo first order kinetics:
dFSUdt
= k '1O2 [1O2]ss (Eq 39.)
Here, k’102 is the pseudo-first order rate constant whose units are FSU uM-1 min-1.
Typically, pseudo-first order and first order rate constants have units of concentration
per concentration per time, which reduces to simply units of per time. In this study, we
are modeling the increase in fluorescence, which does not have units of concentration,
thus we arrive at our slightly modified rate constant. When plotting the observed
increase in fluorescence versus calculated singlet oxygen concentration (Figure 39), the
slope is k’102 = 5.8x109 FSU uM-1 min-1. This value can then be used to convert the
observed rate of fluorescence into the singlet oxygen concentration in solution and is the
result of our experimental conditions, including the excitation and emission
wavelengths of the Modulus fluorimeter used.
156
!
0 1 2 3 4 5 6 70
100
200
300
400
500
600
y=(58.22 ± 4.16)x + (93.80 ± 14.10)
R2 =0.99
ΔFS
U/m
in
[1O2]ss (x10-8 µM)
Figure 39: Change in fluorescence over given steady state singlet oxygen concentrations, determined from a Rose Bengal standard curve. Error bars indicate the standard deviation of triplicate measurements. The slope of this line is k’ as indicated in Equation 39. Error bars indicate the standard deviation of triplicate measurements.
The method for calculating CT, the dose measurement for conventional
disinfectants [275] and cationic fullerenes [33], is shown below
CT = FSU(t)k '1O2
= [1O2 ]0 t (Eq 40.)
where CT is the disinfectant concentration, C, multiplied by the contact time, t. The
disinfectant concentration, C, in the FNP suspensions is taken to be the steady state 1O2
concentration probed by SOSG on or very near the surface of the NP, calculated using
the Rose Bengal experimental standard curve, above. Inactivation data for the FNP
suspensions over 5 minutes is shown in Figure 15 and inactivation versus CT is shown
in Figure 20 (These data are found in Chapter 4.1).
157
Table 7: MS2 inactivation rate constant during control experiments in UV-A. Average rate constant and standard deviation are shown here.
Chemicals added Rate constant (min-1) MS2+ UVA 0.034± 0.001 MS2+C60(NH2)6 + β-carotene+UV-A 0.034 ± 0.003 MS2+C60(OH)6 + β-carotene+UV-A 0.033 ± 0.003 MS2+C60(OH)24 +β-carotene+UV-A 0.035 ± 0.002 MS2+aqu-nC60 + β-carotene+UV-A 0.035 ± 0.004
The results in Table 7 show that in the presence of β-carotene the inactivation
rate is similar to UV-A alone indicating that the inactivation in the absence of β-carotene
is mediated solely by UV-A sensitized 1O2.
Table 8: MS2 viability during control experiments in the dark.
Chemicals added 0 min 5 min MS2 dark (1.75 ± 0.06) x 107 (1.75 ± 0.05) x 107 MS2+aqu-nC60 dark (1.72 ± 0.09) x 107 (1.75 ± 0.06) x 107 MS2+C60(OH)24 dark (1.70 ± 0.07) x 107 (1.69 ± 0.08) x 107 MS2+C60(OH)6 dark (1.72 ± 0.05) x 107 (1.73 ± 0.07) x 107 MS2+C60(NH2)6 dark (1.73 ± 0.07) x 107 (1.76 ± 0.05) x 107
Results in Table 8 show that in the absence of UV-A, fullerenes had no effect on
the viability of MS2.
Hyperspectral image analysis
Figure 40 illustrates the sequence of steps involved in processing a hyperspectral
image of aqu-nC60 and MS2 virus sample. High resolution spectral and spatial
information from hyperspectral images (Figure 40 (a) and (b)) were obtained by first
applying the minimum noise fraction transform to estimate and segregate the noise to
158
generate coherent data (Figure 40 (c) and (d)). Next using two-cascaded principle
component transformations, the noise in the image was converted to unit variance and
the bands were resolved into corresponding standard principal components [374]. The
Eigen values were analyzed using the minimum noise fraction bands (images) to aid in
retrieving the most coherent images (images with Eigen values near 1 are mostly
comprised of noise) (Figure 40 (e)). The coherent image was then analyzed in terms of
pixel purity index (PPI) values computed by repeatedly projecting n-dimensional scatter
plots onto a random unit vector (Figure 40 (f)). The extreme or spectrally “pure” pixels
in each projection (those pixels that fell onto the ends of the unit vector) were recorded
and the PPI image was created by noting the total number of times each pixel was
marked as extreme. Next, the data cloud was rotated using the interactive n-
Dimensional Visualizer (NDVI) algorithm to select, classify, and extract the endmembers
(spectrally pure materials) of the nanoparticles and viruses (Figure 40 (g) and (h)). The
selected endmembers were classified by mixture tuned matched filtering (MTMF)
(Figure 40 (i) and (j)), which mapped the nanoparticles and viruses in the hyperspectral
image. The final step in the analysis is supervised image classification (Figure 40 (k)-(n)).
Similar steps were used for other nanoparticle-virus combinations to determine their
relative association/proximity. Figure 40 (i) and (j) shows the infeasibility vs. MF score
images, where MF-score =1 indicates perfect match. All materials with a score ≥ 0.7 and
infeasibility <20 are chosen as a match to spectral signatures. Figures k-n shows the map
159
of MS2 and aqu-nC60 in noise-free image (Figure 40 (k) and (l)) and hyperspectral image
(Figure 40 (m)-(n)), respectively.
160
8
Figure S6 Hyperspectral image analysis of MS2+aqu-nC60 in aqueous phase. Mixture tuned matched filtering 146 was applied to map the distribution of MS2 and aqu-nC60 in the image. 147
Figure 40: Hyperspectral image analysis of MS2+aqu-nC60 in aqueous phase. Mixture tuned matched filtering was applied to map the distribution of MS2 and aqu-
nC60 in the image.
161
Table 9: Infrared frequencies (wavenumbers) and associated band assignments for MS2 bacteriophage.
Wavenumber (cm-1) Assignment 1733 C=O, carbonyl groups 1710 Glu, ν(C=O) 1695-1670 Intermolecular β-structure 1690-1680 Intramolecular β-structure 1666-1659 ‘3-turn’ helix 1657-1648 α-helix 1645-1640 Random coil 1640-1630 Intramolecular β-structure 1625-1610 Intermolecular β-structure 1606-1 Tyr-O-1 1580 Gln, 1573 HisH, ν(C=C); Asp, νas(COO-) 1550-1510 N-H, secondary amide 1504 C=C, aromatic 1474 δas (CH3) 1449 Pro, ν(C-N) 1405 Glu, νs(COO-); Asp, νs(COO-) 1352-1361 1334-1342
Trp
1420-1181 Ser 1273-6 Trp; Trp-O- 1244 RNA νas(PO2-) 1218 RNA C-H, ring bend 1184 RNA νs(PO2-) 1160 RNA ribose, ν(C-O) 1120 RNA ribose, ν(C-O) 1060 RNA ribose, ν(C-O) 1043 RNA ribose, ν(C-O)
162
10
1800 1750 1700 1650 1600 1550 1500
UV-A
151115
66
1624
1643
1668
1692
(a)MS2
Abso
rban
ce U
nits
(A.U
.)
Wavenumber (cm-1)
1800 1750 1700 1650 1600 1550 1500
UV-A
1710
1733
150415
47
1606
1643
1666
1687
(b)MS2+aqu/nC60
Abso
rban
ce U
nits
(A.U
.)
Wavenumber (cm-1)
1800 1750 1700 1650 1600 1550 1500
UV-A
1580
1616
1637
1678
1657
1688
1733
(c)MS2+ C60(OH)6
Abs
orba
nce
Uni
ts (A
.U.)
Wavenumber (cm-1)
1800 1750 1700 1650 1600 1550 1500
UV-A
1573
1601
1630
1649
1678
1733
(d)MS2+ C60(NH2)6
Abso
rban
ce U
nits
(A.U
.)
Wavenumber (cm-1)
Figure S7. ATR–FTIR spectra of MS2 alone and MS2+fullerenes after exposure UVA for 5 mins. (a) to (d) show the evidence 154 for carbonyl content and protein secondary structures. 155 156 Figure S7 shows the peak fits of amide I band of MS2 in the mid-infrared regions. This region specially 157 represents all the vibrational frequencies associated with protein secondary structures. 158 159 Table S5 Shown here are the location of amino acid (AA) on A-and coat-protein sequences that have potential to react 160 with 1O2 at high second order reaction rates at physiological pH 7.1 [1, 5]. 161 Protein sequence No. of AA targets (A–protein+ coat
protein) for 1O2 ; approx. 2nd order rate constants, M–1s–1 ; neutral pH
A-protein (1 copy/MS2) (GeneBank: ACV72076.1) A-protein+ coat protein 1 mrafstldre netfvpsvrv yadgetedns fslkyrsnwt pgrfnstgak tkqwhypspy 61 srgalsvtsi dqgaykrsgs swgrpyeeka gfgfsldars cyslfpvsqn ltyievpqnv 121 anrastevlq kvtqgnfnlg valaearsta sqlatqtial vkaytaarrg nwrqalryla 181 lnedrkfrsk hvagrwlelq fgwlplmsgi qgayemltkv hlqeflpmra vrqvgtnikl 241 dgrlsypaan fqttcnisrr iviwfyinda rlawlsslgi lnplgivwek vpfsfvvdwl 301 lpvgnmlegl tapvgcsyms gtvtdvitge siisvdapyg wtverqgtak aqisamhrgv 361 qsvwpttgay vkspfsmvht ldalalirqr lsr
5+0: h (Histidine); k= 3.2x107 12+360: w (Tryptophan); k=2-7x107 (physical reaction) k=3x107 (chemical reaction) 16+720: y (Tyrosine); k=0.8x107 8+360: m (Methionine); k=1.6x107 3+360: c (Cysteine); k=8.9x106 All other amino acids: k <0.7x107
Coat protein (180 copies/MS2) (GeneBank: ACY07219.1) 1 masnftqfvl vdnggtgdvt vapsnfangv aewissnsrs qaykvtcsvr qssaqnrkyt 61 ikvevpkvat qtvggvelpv aawrsylnle ltipifatns dcelivkamq gllkdgnpip 121 saiaansgiy
Figure 41: ATR-FTIR spectra of MS2 alone and MS2+fullerenes after UV-A exposure for 5 minutes showing evidence for carbonyl content and protein secondary
structures.
Figure 41 shows the peak fits of amide I band of MS2 in the mid-infrared regions.
This region specially represents all the vibrational frequencies associated with protein
secondary structures.
163
Table 10: Location of amino acids (AA) on A- and coat-protein sequences that have the potential to react with 1O2 at high second order reaction rates at physiological
pH 7.1 [273, 373]
Proteinsequence No. of AA targets (A–protein+ coatprotein) for 1O2 ; approx. 2
nd orderrateconstants,M–1s–1;neutralpH
A-protein(1copy/MS2)(GeneBank:ACV72076.1) A-protein+coatprotein1 mrafstldre netfvpsvrv yadgetedns fslkyrsnwt pgrfnstgaktkqwhypspy61srgalsvtsidqgaykrsgsswgrpyeekagfgfsldarscyslfpvsqnltyievpqnv121anrastevlqkvtqgnfnlgvalaearstasqlatqtialvkaytaarrgnwrqalryla181 lnedrkfrsk hvagrwlelq fgwlplmsgi qgayemltkv hlqeflpmravrqvgtnikl241dgrlsypaanfqttcnisrriviwfyindarlawlsslgilnplgivwekvpfsfvvdwl301 lpvgnmlegl tapvgcsyms gtvtdvitge siisvdapyg wtverqgtakaqisamhrgv361qsvwpttgayvkspfsmvhtldalalirqrlsr
5+0:h(Histidine);k=3.2x10712+360:w(Tryptophan);k=2-7x107(physicalreaction)k=3x107(chemicalreaction)
16+720:y(Tyrosine);k=0.8x1078+360:m(Methionine);k=1.6x1073+360:c(Cysteine);k=8.9x106Allotheraminoacids:k<0.7x107
Coatprotein(180copies/MS2)(GeneBank:ACY07219.1)1 masnftqfvl vdnggtgdvt vapsnfangv aewissnsrs qaykvtcsvrqssaqnrkyt61ikvevpkvatqtvggvelpvaawrsylnleltipifatnsdcelivkamqgllkdgnpip121saiaansgiy
Detailed Procedure of OxyBlot Protein Oxidation Assay.
The protocol provided by the manufacturer was adapted for quantification of
protein oxidation products, specifically carbonyl groups formed by the reactions of 1O2
with amino-acid side chains on MS2. Prior to quantifying the protein carbonyl, the MS2
samples from various treatments (at 5 min) were separately pooled together and
concentrated to get a total protein concentration to between 3–4 µg µL-1. The total
amount of proteins was measured using the Modified Lowry Protein Assay Kit (Pierce
Biotechnology, Rockford, IL) with bovine serum albumin (BSA) standards. The protein
concentration was calculated by measuring the absorbances at 750 nm using UV-vis
spectrophotometer. The procedure is as follows: 5 µl protein samples (3.5 µg µL-1) were
164
first denatured using 5 µl 12% SDS and then mixed with 1x 2,4-dinitrophenylhydrazine
(DNPH) incubated for 15 min during which 2,4-dinitrophenylhydrozone (DNP-
hydrazone) derivative was formed upon reaction with carbonyls. After 15 min, the
reaction was stopped by adding 7.5 µl of neutralization solution. A negative control was
also prepared as described without protein. The derivatized proteins were separated
into individual proteins by 12% sodium dodecyl sulfate polyacrylamide electrophoresis
followed by Western blotting. The protein bands were transferred onto a PVDF
membrane filter and then incubated first with the primary antibody (rabbit anti-DNP
antibody) and then with secondary antibody (goat anti-rabbit IgG (horseradish
peroxidase-conjugated). The filters were then treated with chemiluminescent reagents
(luminol and enhancer). The luminol was converted to the light-emitting form at
wavelength 428 nm by the antigen/primary antibody/secondary antibody/peroxidase
complex (bound to the membrane) in a H2O2 catalyzed oxidation reaction. The light was
detected by short exposure to blue light sensitive films. The oxidative status of each
protein was then analyzed quantitatively by comparing the signal intensity of the same
protein in different lanes on the same gel.
SDS-PAGE.
SDS-PAGE was performed using standard Laemmli gel method [375] to examine
1O2 induced modifications in phage capsid proteins. Pre-cast SDS/12% (w/v)
polyacrylamide gels (Tris-glycine gels) for the BioRad gel apparatus systems used
165
according to the manufacturer’s instructions. The Tris-glycine running buffers and
sample buffers were purchased from Invitrogen (Carlsbad, CA, USA). Phage samples
were diluted by 50% (v/v) in 2x Tris-glycine sample buffer, incubated at 85 oC for 2 min,
and then loaded directly into the wells on the gel. Each well contained 8 µg of protein.
Electrophoresis was carried out for 2-3 h at 30 mA/gel and the gels were stained with
0.1% Coomassie brilliant blue R250 solution according to manufacturer’s instructions
and photographed.
Protein conformational changes are due to the oxidation by 1O2
Figure 42 shows SDS–PAGE analysis of phage proteins obtained from various
treatments: (1) MS2 in dark, (2) MS2+C60(NH2)6 in dark, (3) MS2+UV-A, (4)
MS2+C60(NH2)6+UV-A, (5) MS2+C60(OH)6+UV-A, (6) MS2+aqu-nC60+UV-A, and (7)
MS2+C60(NH2)6+β-carotene+UV-A. MS2 was mainly composed of A-protein (molecular
mass: 44,000 Da) and coat protein (13,700 Da). Similar mobilities of A- and coat-protein
in lanes B and C indicate that C60(NH2)6 in dark had no effect on MS2, which is consistent
with dark inactivation results under these conditions. The A- and coat-protein bands in
lanes D and H are present at slightly higher location (indicates higher molecular mass)
on the gel suggesting that UV-A alone could induce cross–linkages in proteins. Protein
bands in Lanes E (or I), F and G appears to migrate slower when compared with the
bands in Lanes B or C, which is due to oxidation by 1O2. Note that A–protein in Lane E
exhibits a higher molecular mass (> 44 KDa) than any other proteins in other lanes.
166
Similarly, A-proteins in Lanes F and G showed masses > 44 KDa but lower than the mass
of A–protein in Lane E. Coat proteins in Lanes E (or I) F and G showed masses between
14.7–21.5 KDa, which is a consequence of oxidative damage by 1O2.
12
Da). Similar mobilities of A- and coat-protein in lanes B and C indicate that C60(NH2)6 in dark had no effect on 195 MS2, which is consistent with dark inactivation results under these conditions. The A- and coat-protein bands 196 in lanes D and H are present at slightly higher location (indicates higher molecular mass) on the gel suggesting 197 that UVA alone could induce cross–linkages in proteins. Protein bands in Lanes E (or I), F and G appears to 198 migrate slower when compared with the bands in Lanes B or C, which is due to oxidation by 1O2. Note that A–199 protein in Lane E exhibits a higher molecular mass (> 44 KDa) than any other proteins in other lanes. Similarly, 200 A-proteins in Lanes F and G showed masses > 44 KDa but lower than the mass of A–protein in Lane E. Coat 201 proteins in Lanes E (or I) F and G showed masses between 14.7–21.5 KDa, which is a consequence of oxidative 202 damage by 1O2. 203 204
205 Figure S8. Analysis by SDS–PAGE showing the altered migrations of MS2 capsid proteins (A-protein and coat protein) 206 following exposure to UVA sensitized 1O2. Lane A: molecular mass markers; Lane B: MS2 in dark; Lane C: MS2+C60(NH2)6 in 207 dark; Lane D: MS2+UVA; Lane E: MS2+C60(NH2)6+UVA; Lane F: MS2+C60(OH)6+UVA; Lane G: MS2+aqu-nC60+UVA; Lane H: 208 MS2+C60(NH2)6+β-carotene+UVA; Lane I: MS2+C60(NH2)6+UVA 209
210
References: 211 1. Wilkinson, F.; Helman, W. P.; Ross, A. B., Rate constants for the decay and reactions of the lowest 212
electronically excited state of molecular oxygen in solution. An expanded and revised compliation. Journal 213 of Physical and Chemical Reference Data 1995, 34, 663-1021. 214
2. Crittenden, J. C.; Trussell, R. R.; Hand, D. W.; Howe, K. J.; Tchobanoglous, G., Water Treatment: Principles 215 and Design. 2 ed.; John Wiley & Sons, Inc.: Hoboken, NJ, 2005; p 1948. 216
3. Cho, M.; Lee, J.; Mackeyev, Y.; Wilson, L. J.; Alvarez, P. J. J.; Hughes, J. B.; Kim, J.-H., Visible Light Sensitized 217 Inactivation of MS-2 Bacteriophage by a Cationic Amine-Functionalized C60 Derivative. Environ. Sci. 218 Technol. 2010, 44, (17), 6685-6691. 219
4. Green, A. A.; Berman, M.; Switzer, P.; Craig, M. D., A Transformation for ordering multispectral data in 220 terms of image quality with implications for noise removal. IEEE Trans. Geosci. Rem. Sen. 1988, 26, (1), 65-221 74. 222
Figure 42: Analysis by SDS-PAGE showing the altered migrations of MS2 capsid proteins (A-protein and coat protein) following exposure to UV-A sensitized
1O2. Lane A: molecular mass markers; Lane B: MS2 in Dark; Lane C: MS2+C60(NH2)6 in dark; Lane D: MS2+UV-A; Lane E: MS2+ C60(NH2)6+UV-A; Lane F: MS2+C60(OH)6+UV-A; Lane G: MS2+aqu-nC60+UV-A; Lane H: MS2+ C60(NH2)6+β-carotene+UV-A; Lane I:
MS2+ C60(NH2)6+UV-A
167
Appendix B – Supporting information for Chapter 4.2: Influence of Inorganic Anions on the Reactivity of TiO2 Nanoparticles Radiation intensity as a function of emission wavelength at the upper liquid surface
The emission spectrum of the UV lamps, as provided by the manufacturer, was
used to determine the radiation intensity as a function of the emission wavelength at the
liquid upper surface of the samples. Results are reported in Figure 43. A normal
distribution with mean 365 nm and standard deviation 10 nm was assumed for Philips
TL-D 15W BLB SLV spectrum. Radiometric readings were corrected by taking into
account the spectral response of the sensor, as suggested in [376].
2
deviation 10 nm was assumed for Philips TL-D 15W BLB SLV spectrum. Radiometric readings were 1
corrected by taking into account the spectral response of the sensor, as suggested in [1]. 2
3
4
Figure SI1 Radiation intensity vs. emission wavelength at the liquid upper surface of samples. 5
6
SI2. Characterization of size distribution of TiO2 aggregates as a function of operating conditions 7
8
(a) (b)
(c) (d)
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
300 310 320 330 340 350 360 370 380 390
I (m
W c
m-2
)
λ (nm)
10−2 10−1 100 101 102 1030
5
10
15
size (µm)
Inte
nsity
DI t0DI t30KI t0KI t30
10−2 10−1 100 101 102 1030
1
2
3
4
5
6
7
8
9
10
size (µm)
Inte
nsity
DI t0DI t30KI t0KI t30
Figure 43: Radiation intensity vs. emission wavelength at the upper liquid surface of samples.
Characterization of size distribution of TiO2 aggregates as a function of operating
conditions
168
3
(a) (b)
(c) (d)
Figure SI2 TiO2 aggregate size distribution as a function of KI concentration (0, 50 mM) at various times 1
(0, 30 min) (a,b). TiO2 aggregate D50 versus time as a function of KI concentration (0,50mM) (c,d). 2
Experimental results are expressed both as particle number weighted (a,c) and volume weighted (b,d). 3
4
SI3. Aggregation of TiO2 NPs in inorganic anions 5
6
(a) (b)
10−2 10−1 100 101 102 1030
5
10
15
size (µm)
Inte
nsity
DI t0DI t30KI t0KI t30
10−2 10−1 100 101 102 1030
1
2
3
4
5
6
7
8
9
10
size (µm)
Inte
nsity
DI t0DI t30KI t0KI t30
0 5 10 15 20 25 300
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
time (min)
D50
(µm
)
DI50mM KI
0 5 10 15 20 25 300
5
10
15
20
25
30
35
40
time (min)
D50
(µm
)
DI50mM KI
Figure 44: TiO2 aggregate size as a function of KI concentration (0, 50 mM) at various times (0, 30 min) (a,b). TiO2 aggregate D50 vs. time as a function of KI
concentration (0, 50 mM) (c, d). Experimental results are expressed both as number weighted (a, c) and volume weighted (b, d).
Aggregation of TiO2 NPs in inorganic anions
169
3
Figure SI2 TiO2 aggregate size distribution as a function of KI concentration (0, 50 mM) at various times 1
(0, 30 min) (a,b). TiO2 aggregate D50 versus time as a function of KI concentration (0,50mM) (c,d). 2
Experimental results are expressed both as particle number weighted (a,c) and volume weighted (b,d). 3
4
SI3. Aggregation of TiO2 NPs in inorganic anions 5
6
(a) (b)
Figure SI3.1 TiO2 aggregate size over 30minutes as a function of anion at 0.5mM, number weighted D50 (a) and 7
volume weighted D50 (b). 8
(a) - Carbonate (b)
0 5 10 15 20 25 300
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
time (min)
D50
(µm
)
DI50mM KI
0 5 10 15 20 25 300
5
10
15
20
25
30
35
40
time (min)
D50
(µm
)
DI50mM KI
0 5 10 15 20 25 300
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
time (min)
D50
(µm
)
PhosphateChlorideNitrateCarbonateSulfate
0 5 10 15 20 25 30−10
0
10
20
30
40
50
60
time (min)
D50
(µm
)
PhosphateChlorideNitrateCarbonateSulfate
Figure 45: TiO2 aggregate size over 30 minutes as a function of anion at 0.5mM, (a) number weighted D50, and (b) volume weighted D50.
170 5
(a) - Carbonate (b)
(c) - Phosphate (d)
(e) - Nitrate (f)
0 5 10 15 20 25 30
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
0 5 10 15 20 25 300
10
20
30
40
50
60
70
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
0 5 10 15 20 25 30
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
0 5 10 15 20 25 300
10
20
30
40
50
60
70
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
0 5 10 15 20 25 30
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
0 5 10 15 20 25 300
10
20
30
40
50
60
70
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
171 6
(g) - Chloride (h)
(i) - Sulfate (j)
Figure SI3.2 TiO2 aggregate size over 30minutes as a function of anion concentration, number weighted D50 1
(a,c,e,g,i) and volume weighted D50 (b,d,f,h,j). 2
3
(a) (b)
0 5 10 15 20 25 30
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
0 5 10 15 20 25 300
10
20
30
40
50
60
70
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
0 5 10 15 20 25 30
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
0 5 10 15 20 25 300
10
20
30
40
50
60
70
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
0 2 4 6 8 10 12 140
5
10
15
20
25
30
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
0 2 4 6 8 10 12 140
5
10
15
20
25
30
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
Figure 46: TiO2 aggregate size over 30 minutes as a function of anion concentration, (a, c, e, g, i) number weighted D50 and (b, d, f, h, j) volume weighted
D50.
172
5
(i) - Sulfate (j)
Figure SI3.2 TiO2 aggregate size over 30minutes as a function of anion concentration, number weighted D50 1
(a,c,e,g,i) and volume weighted D50 (b,d,f,h,j). 2
3
(a) (b)
0 5 10 15 20 25 30
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
0 5 10 15 20 25 300
10
20
30
40
50
60
70
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
0 5 10 15 20 25 30
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
0 5 10 15 20 25 300
10
20
30
40
50
60
70
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
0 2 4 6 8 10 12 140
5
10
15
20
25
30
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
0 2 4 6 8 10 12 140
5
10
15
20
25
30
time (min)
D50
(µm
)
0.5mM2.5mM5mM12.5mM25mM
Figure 47: TiO2 aggregate size over 14 minutes as a function of anion concentration, volume weighted D50 for (a) nitrate and (b) sulfate.
Gouy Chapman Calculations
DLVO potential energy curves were calculated using the Gouy Chapman
approximation for two identical spheres, in which electrostatic repulsion, Φr, and Van
der Waals attraction, Φa, can be calculated as follows:
φr =32πεkB
2T 2γ02a
e2z2e−κx (Eq 41.)
and
φa = −A6
2s2 − 4
−2s2+ ln s2 − 4
s2"
#$
%
&'
(
)*
+
,- (Eq 42.)
Where
s = 2a+ xa
(Eq 43.)
and
173
γ0 =exp 2eψ0
2kBT!
"#
$
%&−1
exp 2eψ0
2kBT!
"#
$
%&+1
(Eq 44.)
Here, ε is the dialectric constant (ε = ε0*εr), kB is the boltzman constant, T is
temperature, a is the radius of the particle, x is the separation distance, A is the
hammaker constant for TiO2 (A= 60 x 10-21 J), e is the charge of an electron, I is the ionic
strength, Ψ0 is the surface potential (assumed to be equal to the measured zeta potential),
and κ is the inverse debeye length, calculated as
κ −1 =εkBT2NAe
2I≈0.304
I (Eq 45.)
Calculations were performed for TiO2 in the 1:1 electrolyte KCl from 0.5 to 25
mM. As the ionic strength increases, the zeta potential at pH 7.8 diminishes in
magnitude (see Figure 50 (c), and compression of the double layer is observed, resulting
in wholly negative potential energies for the highest two concentrations. Ionic strengths
for which the zeta potential was not measured were estimated from available data.
174 8
(a) (b)
Figure SI4.1 Potential energy curves of DLVO theory using the Gouy Chapman approximation for 20 mg L-1 TiO2 1
as a function of KCl concentration. Blue dashed lines are electrostatic repulsion, red dashed lines are Van der 2
Waals attraction, and black solid lines are the summation of repulsive and attractive forces for a given ionic 3
strength. Full plot (a) and detail (b). 4
5
SI5. Fractal Dimension of TiO2 NPs 6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
0 5 10 15 20 25 30 35 40 45 50−2.5
−2
−1.5
−1
−0.5
0
0.5
1x 10−18
Separation Distance (nm)
Pote
ntia
l Ene
rgy
(J)
0 5 10 15 20 25 30 35 40 45 50
−0.8
−0.6
−0.4
−0.2
0
0.2
0.4
0.6
0.8
x 10−20
Separation Distance (nm)
Pote
ntia
l Ene
rgy
(J)
Increasing*ionic*strength*
Figure 48: Potential energy curves of DLVO theory using the Gouy Chapman approximation for 20 mg L-1 TiO2 as a function of KCl concentration (0.5 – 25 mM). (a) Full plot and (b) detail. Blue dashed lines are electrostatic repulsion, red dashed lines
are Van der Waals attraction, and black solid lines are the summation of repulsive and attractive forces for a given ionic strength.
Fractal Dimension of TiO2 NPs
175 9
(a) – 0.5mM Carbonate (b)
(c) – 25mM Carbonate (d)
(e) – 0.5mM Phosphate (f)
−3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1−6
−5
−4
−3
−2
−1
0
Log(I) vs Log(q)Linear Regression (Solid Blue), Measured Values (Dashed Black)
log(q)
log(
I)
0.65 0.7 0.75 0.8 0.85 0.9 0.95 1−3.8
−3.6
−3.4
−3.2
−3
−2.8
−2.6
−2.4Linear Regression (Blue) compared to Measured Values (circles)
log(q)
log(
I)
−3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1−4.5
−4
−3.5
−3
−2.5
−2
−1.5
−1
−0.5
0
Log(I) vs Log(q)Linear Regression (Solid Blue), Measured Values (Dashed Black)
log(q)
log(
I)
−0.4 −0.2 0 0.2 0.4 0.6 0.8 1−4.5
−4
−3.5
−3
−2.5
−2
−1.5
−1Linear Regression (Blue) compared to Measured Values (circles)
log(q)
log(
I)
−3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1−6
−5
−4
−3
−2
−1
0
Log(I) vs Log(q)Linear Regression (Solid Blue), Measured Values (Dashed Black)
log(q)
log(
I)
0.65 0.7 0.75 0.8 0.85 0.9 0.95 1−6
−5.5
−5
−4.5
−4
−3.5
−3Linear Regression (Blue) compared to Measured Values (circles)
log(q)
log(
I)
176 10
(g) – 25mM Phosphate (h)
(i) – 0.5mM Chloride (j)
(k) – 25mM Chloride (l)
−3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1−5
−4.5
−4
−3.5
−3
−2.5
−2
−1.5
−1
−0.5
0
Log(I) vs Log(q)Linear Regression (Solid Blue), Measured Values (Dashed Black)
log(q)
log(
I)
0.65 0.7 0.75 0.8 0.85 0.9 0.95 1−5
−4.5
−4
−3.5
−3
−2.5
−2Linear Regression (Blue) compared to Measured Values (circles)
log(q)
log(
I)
−3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1−5
−4.5
−4
−3.5
−3
−2.5
−2
−1.5
−1
−0.5
0
Log(I) vs Log(q)Linear Regression (Solid Blue), Measured Values (Dashed Black)
log(q)
log(
I)
−0.4 −0.2 0 0.2 0.4 0.6 0.8 1−5
−4.5
−4
−3.5
−3
−2.5
−2
−1.5Linear Regression (Blue) compared to Measured Values (circles)
log(q)
log(
I)
−3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1−6
−5
−4
−3
−2
−1
0
Log(I) vs Log(q)Linear Regression (Solid Blue), Measured Values (Dashed Black)
log(q)
log(
I)
−0.4 −0.2 0 0.2 0.4 0.6 0.8 1−5.5
−5
−4.5
−4
−3.5
−3
−2.5
−2Linear Regression (Blue) compared to Measured Values (circles)
log(q)
log(
I)
177 11
(i) – 0.5mM Nitrate (j)
(i) – 25mM Nitrate (j)
(i) – 0.5mM Sulfate (j)
−3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1−6
−5
−4
−3
−2
−1
0
Log(I) vs Log(q)Linear Regression (Solid Blue), Measured Values (Dashed Black)
log(q)
log(
I)
−0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1−5.5
−5
−4.5
−4
−3.5
−3
−2.5
−2
−1.5
−1Linear Regression (Blue) compared to Measured Values (circles)
log(q)
log(
I)
−3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1−5
−4.5
−4
−3.5
−3
−2.5
−2
−1.5
−1
−0.5
0
Log(I) vs Log(q)Linear Regression (Solid Blue), Measured Values (Dashed Black)
log(q)
log(
I)
−0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1−5
−4.5
−4
−3.5
−3
−2.5
−2
−1.5
−1Linear Regression (Blue) compared to Measured Values (circles)
log(q)
log(
I)
−3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1−5
−4.5
−4
−3.5
−3
−2.5
−2
−1.5
−1
−0.5
0
Log(I) vs Log(q)Linear Regression (Solid Blue), Measured Values (Dashed Black)
log(q)
log(
I)
−0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1−5
−4.5
−4
−3.5
−3
−2.5
−2
−1.5
−1
−0.5
0Linear Regression (Blue) compared to Measured Values (circles)
log(q)
log(
I)
178 12
(i) – 25mM Sulfate (j)
Figure SI5 – Representative SLS data and calculated fits at 0 and 30 minutes for each anion. 1
2
SI6. Zeta potential versus pH as a function of anion concentration 3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
−3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1−6
−5
−4
−3
−2
−1
0
Log(I) vs Log(q)Linear Regression (Solid Blue), Measured Values (Dashed Black)
log(q)
log(
I)
−0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1−5.5
−5
−4.5
−4
−3.5
−3
−2.5
−2Linear Regression (Blue) compared to Measured Values (circles)
log(q)
log(
I)
Figure 49: Representative SLS data and calculated fits at 0 and 30 minutes for each anion.
Zeta potential versus pH as a function of anion concentration
179 11
(a) - Carbonate (b) - Phosphate
(c) - Chloride (d) - Nitrate
(e) - Sulfate
1
SI7. Influence of initial iodide concentration on hole detection 2
2 3 4 5 6 7 8 9 10−50
−40
−30
−20
−10
0
10
20
30
40
50
pH
ZP (m
V)
0.005mM0.5mM0.5mM5mM12.5mM
2 3 4 5 6 7 8 9 10−50
−40
−30
−20
−10
0
10
20
30
40
50
pH
ZP (m
V)
0.005mM0.5mM0.5mM5mM12.5mM
2 3 4 5 6 7 8 9 10−50
−40
−30
−20
−10
0
10
20
30
40
50
pH
ZP (m
V)
0.5mM5mM12.5mM
2 3 4 5 6 7 8 9 10−50
−40
−30
−20
−10
0
10
20
30
40
50
pH
ZP (m
V)
0.5mM5mM12.5mM
2 3 4 5 6 7 8 9 10−50
−40
−30
−20
−10
0
10
20
30
40
50
pH
ZP (m
V)
0.005mM0.5mM0.5mM5mM12.5mM
Figure 50: Zeta potential vs. pH as a function of anion concentration
Influence of initial iodide concentration on hole detection
180 12
(a) (b)
Figure SI7 Absorption from iodo-starch method versus time as a function of initial iodide concentration (a) and 1
photogenerated hole detection (measured as oxidized iodide concentration) after 30 minutes irradiation 2
versus initial iodide concentration (b). 3
4
SI8. Time series for iodide oxidation as a function of inorganic anion concentration 5
(a) (b)
Figure SI8 Oxidized iodide concentration (mean±st.dev.) vs. time as a function of inorganic anion 6
concentration: carbonate (a) and phosphate (b). 7
8
SI9. Langmuir-Hinshelwood linearization procedure for KA and kR constants estimation 9
By neglecting the presence of inorganic anions other than iodide, equation 3 in the manuscript 10
can be simplified and expressed showing the linear relationship between the inverse of initial 11
reaction rate and the inverse of initial iodide concentration, as follows [2]: 12
0 5 10 15 20 25 300
0.05
0.1
0.15
0.2
0.25
0.3
0.35
Time(min)
ABS
25mM37.5mM50mM75mM100mM
20 40 60 80 100 1200
0.005
0.01
0.015
0.02
0.025
0.03
Iodide (mM)
I ox (m
M)
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0 5 10 15 20 25 30
[I-O
X] (
mM
)
Time (min)
1 mM 2.5 mM 5 mM 10 mM 25 mM
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0 5 10 15 20 25 30
[I-O
X] (
mM
)
Time (min)
1 mM 2.5 mM 5 mM 10 mM 25 mM
Figure 51: (a) Absorption from iodo-starch method vs. time as a function of initial iodide concentration and (b) photogenerated hole detection (measured as
oxidized iodide concentration) after 30 minutes irradiation versus initial concentration.
Time series for iodide oxidation as a function of inorganic anion concentration
12
(a) (b)
Figure SI7 Absorption from iodo-starch method versus time as a function of initial iodide concentration (a) and 1
photogenerated hole detection (measured as oxidized iodide concentration) after 30 minutes irradiation 2
versus initial iodide concentration (b). 3
4
SI8. Time series for iodide oxidation as a function of inorganic anion concentration 5
(a) (b)
Figure SI8 Oxidized iodide concentration (mean±st.dev.) vs. time as a function of inorganic anion 6
concentration: carbonate (a) and phosphate (b). 7
8
SI9. Langmuir-Hinshelwood linearization procedure for KA and kR constants estimation 9
By neglecting the presence of inorganic anions other than iodide, equation 3 in the manuscript 10
can be simplified and expressed showing the linear relationship between the inverse of initial 11
reaction rate and the inverse of initial iodide concentration, as follows [2]: 12
0 5 10 15 20 25 300
0.05
0.1
0.15
0.2
0.25
0.3
0.35
Time(min)
ABS
25mM37.5mM50mM75mM100mM
20 40 60 80 100 1200
0.005
0.01
0.015
0.02
0.025
0.03
Iodide (mM)
I ox (m
M)
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0 5 10 15 20 25 30
[I-O
X] (
mM
)
Time (min)
1 mM 2.5 mM 5 mM 10 mM 25 mM
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0 5 10 15 20 25 30
[I-O
X] (
mM
)
Time (min)
1 mM 2.5 mM 5 mM 10 mM 25 mM
Figure 52: Oxidized iodide concentration (mean±std) vs. time as a function of inorganic anion concentration: (a) carbonate and (b) phosphate.
Langmuir-Hinshelwood linearization procedure for KA and kR constants estimation
By neglecting the presence of inorganic anions other than iodide, Equation 33 in
Chapter 4.2 can be simplified and expressed showing the linear relationship between the
181
inverse of initial reaction rate and the inverse of initial iodide concentration, as follows
[313]:
1r0=
1K
A,I−KR
1[I − ]0
+1KR
=m 1[I − ]0
+ q (Eq 46.)
With
KR =1q
(Eq 47.)
and
KA =qm
(Eq 48.)
The rate r0 is obtained from experimental data by calculating the oxidized iodide
concentration for each [I-]0 over time. KA and KR values were calculated by fitting
experimental data shown with a linear regression, estimating m and q values by the
least squares method, as in Figure 53.
13
1!!!= ! 1
!!,!−!!!! 1!! !
+ ! 1!!!= !!! 1
!! !!+ !! (e.SI1)
with 1
!! = ! !! !! = !! 2
The rate r0 is obtained from experimental data by calculating the oxidized iodide concentration for 3
each [I-]0 over time. KA and KR values were calculated by fitting experimental data shown in figure 2 4
with a linear regression, estimating m and q values by least square method, as in figure SI5. 5
6
7
Figure SI9.1 Experimental data linear regressions (1/r vs. 1/[I-]0) for iodide adsorption via Langmuir-Hinshelwood 8
mechanism for sonicated TiO2. 9
10
Equation 3 in the manuscript can be transformed so as to highlight the linear relationship between 11
the inverse of the initial reaction rate and the initial inorganic anion concentration, as it follows [3]: 12
1!!!= ! !!,!"#$"!
!!!!,!−! !! !![!"#$"]! + !
1 + !!,!−! !! !!!!!,!−! !! !
!= !!![!"#$"]! !+ !! (e.SI2)
4.0E+04
6.0E+04
8.0E+04
1.0E+05
1.2E+05
1.4E+05
1.6E+05
1.8E+05
2.0E+05
0.00 0.01 0.02 0.03 0.04 0.05
1/r (
s)
1/[I-]0 (mM)
Iodide, S
Figure 53: Experimental data linear regressions (1/r vs 1/[I-]0) for iodide adsorption via Langmuir-Hinshelwood mechanism for sonicated TiO2.
182
Equation 33 in Chapter 4.2 can be transformed so as to highlight the linear
relationship between the inverse of the initial reaction rate and the initial inorganic
anion concentration, as it follows [286]:
1r0=
KA,Anion
KA,I−KR[I
− ]0[Anion]0 +
1+KA,I−[I − ]0
KA,I−KR[I
− ]0=m[Anion]0 + q (Eq 49.)
With
KR =1+K
A,I−[I − ]0
qKA,I−[I − ]0
(Eq 50.)
and
KA =m ⋅KRKA,I−[I − ]0 (Eq 51.)
Experimental values at 0 and 12.5 mM were neglected in calculations because of
the modifications on TiO2 suspension characteristics. Results for linearization of
experimental data by the proposed procedure are reported in Figure 54.
183
14
with 1
!! = !!!!!,!−! !! !!⋅!!,!−! !! !
!! = ! ⋅ !!!!,!−! !! ! 2
In calculation procedure experimental values at 0 and 12.5 mM were neglected because of the 3
modifications on TiO2 suspension characteristics. Results for linearization of experimental data by the 4
proposed procedure are reported in figure SI6. 5
6
(a) 7
8
(b) 9
10
Figure SI9.2 Experimental data linear regressions (1/r vs. [Anion]) for the competitive adsorption of inorganic 11
anions with iodide via Langmuir-Hinshelwood mechanism for various anionic species, i.e. carbonate (a) and 12
phosphate (b). 13
6.0E+04
6.5E+04
7.0E+04
7.5E+04
8.0E+04
8.5E+04
9.0E+04
0 1 2 3 4 5
1/r (
s)
[Anion] (mM)
Carbonate, S+A
5.0E+04
1.0E+05
1.5E+05
2.0E+05
2.5E+05
3.0E+05
0 1 2 3 4 5
1/r (
s)
[Anion] (mM)
Phosphate, S+A
Figure 54: Experimental data linear regressions (1/r vs. [Anion]) for the competitive adsorption of inorganic anions with iodide via Langmuir Hinshelwood
mechanism for (a) carbonate and (b) phosphate.
Time Series for terephthalic acid oxidation as a function of inorganic anion
concentration
184
15
SI10. Time series for terephthalic acid oxidation as a function of inorganic anion concentration 1
(a) (b)
(c) (d)
(e)
Figure SI10 2-hydroxyterephthalic acid concentration (mean±st.dev.) vs. time as a function of inorganic anion 2
concentration for various anionic species, i.e. carbonate (a), chloride (b), nitrate (c), sulfate (d) and 3
phosphate (e). 4
0.000
0.004
0.008
0.012
0.016
0 5 10 15 20 25 30
[2-H
TA] (
mM
)
Time (min)
0 mM 1 mM 5 mM 10 mM 25 mM
0.000
0.004
0.008
0.012
0.016
0 5 10 15 20 25 30
[2-H
TA] (
mM
)
Time (min)
0 mM 1 mM 5 mM 10 mM 25 mM
0.000
0.004
0.008
0.012
0.016
0 5 10 15 20 25 30
[2-H
TA] (
mM
)
Time (min)
0 mM 1 mM 5 mM 10 mM 25 mM
0.000
0.004
0.008
0.012
0.016
0 5 10 15 20 25 30
[2-H
TA] (
mM
)
Time (min)
0 mM 1 mM 5 mM 10 mM 25 mM
0.000
0.004
0.008
0.012
0.016
0 5 10 15 20 25 30
[2-H
TA] (
mM
)
Time (min)
0 mM 1 mM 5 mM 10 mM 25 mM
Figure 55: 2-Hydroxyterephthalic acid concentration (mean±std) vs. time as a function of inorganic anion concentration for various anionic species: (a) carbonate,
(b) chloride, (c)nitrate, (d) sulfate, and (e) phosphate.
Normalized hydroxyl radical generation as a function of inorganic anion
concentration.
185
16
SI11. Normalized hydroxyl radical generation as a function of inorganic anion concentration 1
2
3
Figure SI11 Normalized hydroxyl radical generation vs. inorganic anion concentration after 30 minute 4
irradiation time as a function of anionic species in solution. 5
6
REFERENCES 7
[1] Grcic I., Li Puma G., Photocatalytic degradation of water contaminants in multiple 8
photoreactors and evaluation of reaction kinetic constants independent of photon absorption, 9
irradiance, reactor geometry, and hydrodynamics, Environ. Sci. Technol. 47 (2013) 13702-13711. 10
[2] Herrmann J.M., Pichat P., Oxidation of halide ions by oxygen in ultraviolet irradiated aqueous 11
suspension of titanium dioxide. J.C.S. Faraday I 76 (1980) 1138-1146. 12
[3] Chen H.Y., Zahraa O., Bouchy M., Inhibition of the adsorption and photocatalytic degradation 13
of an organic contaminant in an aqueous suspension of TiO2 by inorganic ions. J. Photochem. 14
Photobiol. A 108 (1997) 37-44. 15
16
0.0
0.2
0.4
0.6
0.8
1.0
0 5 10 15 20 25
No
rma
lize
d h
ydro
xyl r
ad
ica
l g
ene
ratio
n
[Anion] (mM)
Carbonate Chloride Nitrate Phosphate Sulfate
Figure 56: Normalized hydroxyl radical generation vs. inorganic anion concentration after 30 minute irradiation time as a function of anionic species in
solution.
186
Appendix C – Supporting Information for Chapter 4.3: Chlorpyrifos Degradation via Photoreactive TiO2 Nanoparticles: Assessing the Impact of a Multi-Component Degradation Scenario TiO2 Characterization and Reactivity 1. TiO2 Characterization and Reactivity
(a) (b)
(c) (d)
Figure SI-1 - TiO2 Characterization a) Time resolved DLS b) TEM image of TiO2 aggregate in MD media; scale bar is 50nm c) DLS of TiO2 in MD media, unweighted, d) ZP vs pH of 20ppm TiO2 in MD, adjusted with HCl. EPM measurements for TiO2 in MD
titrated with 0.1N HCl indicate that the IEP occurs at a pH of roughly 1.5.
0 50 100 150 200 250 300 3500
20
40
60
80
100
120TiO2 Stability in MD, Number Weighted
Time (min)
Rad
ius
(nm
)
Mean = 91.9±2.7nm
100 101 102 1030
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Radius (nm)
Unw
eigh
ted
Inte
nsity
1 2 3 4 5 6 7 8 9−60
−50
−40
−30
−20
−10
0
pH
ZP (m
V)
IEP ~ 1.5
Figure 57: TiO2 Characterization: (a) time resolved DLS, (b) TEM image of TiO2 aggregate in MD media; scale bar is 50nm, (c) DLS of TiO2 in MD media, unweighted,
(d) ZP vs. pH of 20 ppm TiO2 in MD, adjusted with HCl.
187
Figure SI-2 – Reactivity measurements of Hydroxyl Radical and Superoxide from 20ppm TiO2 in MD exposed to UV, in the presence of quenching agents, and in the dark.
2. Stability of TiO2 in CPF
The presence of CPF was not observed to induce TiO2 aggregation. Time resolved DLS was performed over 200 minutes during which time the aggregate size remained constant.
Figure SI-3 – Number weighted TRDLS of 20ppm TiO2 in presence of 385ppb CPF
3. CPF UV/TiO2 Degradation
TiO2 TiO2+Quencher Dark0
0.2
0.4
0.6
0.8
1
1.2
Rela
tive
Fluo
resc
ence
(FSU
)
Hydroxyl RadicalSuperoxide
0 50 100 150 2000
20
40
60
80
100
120TiO2 Stability with CPF, Number Weighted
Time (min)
Rad
ius
(nm
)
Mean = 91.2±1.4nm
Figure 58: Reactivity measurements of hydroxyl radical and superoxide from 20 ppm TiO2 in MD exposed to UV, exposed to UV in the presence of quenching
agents, and in the dark.
Stability of TiO2 in CPF
The presence of CPF was not observed to induce TiO2 aggregation. Time
resolved DLS was performed over 200 minutes during which time the aggregate size
remained constant.
Figure SI-2 – Reactivity measurements of Hydroxyl Radical and Superoxide from 20ppm TiO2 in MD exposed to UV, in the presence of quenching agents, and in the dark.
2. Stability of TiO2 in CPF
The presence of CPF was not observed to induce TiO2 aggregation. Time resolved DLS was performed over 200 minutes during which time the aggregate size remained constant.
Figure SI-3 – Number weighted TRDLS of 20ppm TiO2 in presence of 385ppb CPF
3. CPF UV/TiO2 Degradation
TiO2 TiO2+Quencher Dark0
0.2
0.4
0.6
0.8
1
1.2
Rela
tive
Fluo
resc
ence
(FSU
)
Hydroxyl RadicalSuperoxide
0 50 100 150 2000
20
40
60
80
100
120TiO2 Stability with CPF, Number Weighted
Time (min)
Rad
ius
(nm
)
Mean = 91.2±1.4nm
Figure 59: Number weighted TRDLS of 20 ppm TiO2 in the presence of 385 ppb CPF.
188
CPF UV/TiO2 Degradation
CPF can be susceptible to base catalyzed hydrolysis, however no degradation
products were observed for any of the dark series. Additionally, the working stock
solutions showed no decrease in concentration over time.
CPF can be susceptible to base catalyzed hydrolysis, however no degradation products were observed for any of the dark series. Additionally, the working stock solutions showed no decrease in concentration over time.
(a) (b)
Figure SI-4 – a) CPF, CPF oxon, and TCP over time in the dark with 0ppm TiO2 (grey
icons) and 20ppm TiO2 (black icons) b) Relative working stock concentrations measured at 0, 60 minutes in the dark
4. Bacterial Inactivation in the Presence of CPF
(a) (b)
Figure SI-5 – Bacterial inactivation due to 20ppm TiO2 exposed to UV in the presence and absence of CPF. a) B. subtilis and b) A. baumanii.
5. CPF Degradation using CAKE Software
0 200 400 600 800 1000 1200 14000
50
100
150
200
250
300
Time (min)
ng/m
l
CPF CPF Oxon TCP0min 60min
0
0.2
0.4
0.6
0.8
1
Rel
ativ
e C
once
ntra
tion
0 10 20 30 40 50 600
0.2
0.4
0.6
0.8
1
1.2
1.4B.subtilis Inactivation
Time (min)
N/N
0
UV+, CPF UV+, no CPF
0 10 20 30 40 50 600
0.2
0.4
0.6
0.8
1
1.2
1.4A.baumannii Inactivation
Time (min)
N/N
0
UV+, CPF UV+, no CPF
Figure 60: (a) CPF, CPF oxon, and TCP over time in the dark with 0 ppm TiO2 (grey icons) and 20 ppm TiO2 (black icons). (b) Relative working stock concentrations
measured at 0, 60 minutes in the dark.
Bacterial Inactivation in the Presence of CPF
189
CPF can be susceptible to base catalyzed hydrolysis, however no degradation products were observed for any of the dark series. Additionally, the working stock solutions showed no decrease in concentration over time.
(a) (b)
Figure SI-4 – a) CPF, CPF oxon, and TCP over time in the dark with 0ppm TiO2 (grey
icons) and 20ppm TiO2 (black icons) b) Relative working stock concentrations measured at 0, 60 minutes in the dark
4. Bacterial Inactivation in the Presence of CPF
(a) (b)
Figure SI-5 – Bacterial inactivation due to 20ppm TiO2 exposed to UV in the presence and absence of CPF. a) B. subtilis and b) A. baumanii.
5. CPF Degradation using CAKE Software
0 200 400 600 800 1000 1200 14000
50
100
150
200
250
300
Time (min)
ng/m
l
CPF CPF Oxon TCP0min 60min
0
0.2
0.4
0.6
0.8
1
Rel
ativ
e C
once
ntra
tion
0 10 20 30 40 50 600
0.2
0.4
0.6
0.8
1
1.2
1.4B.subtilis Inactivation
Time (min)
N/N
0
UV+, CPF UV+, no CPF
0 10 20 30 40 50 600
0.2
0.4
0.6
0.8
1
1.2
1.4A.baumannii Inactivation
Time (min)
N/N
0
UV+, CPF UV+, no CPF
Figure 61: Bacterial inactivation due to 20 ppm TiO2 exposed to UV in the presence and absence of CPF (a) B. subtilis and (b) A. baumanii.
CPF Degradation using CAKE Software
Figure SI-6 – CAKE software fit CPF, CPF Oxon, and TCP concentration versus
time. Calculated half-lives are presented in the inset.
The following data are reported from the CAKE software package using single first order (SFO) kinetic fits of experimental CPF, CPF Oxon and TCP data. Initial Values of Sequence Parameters:
Parameter Initial Value Bounds Fixed Parent_0 781 0 to (unbounded) No k_Parent 0.1 0 to (unbounded) No f_Parent_to_A1 0.3333 0 to 1 No f_Parent_to_B1 0.3333 0 to 1 No k_A1 0.1 0 to (unbounded) No f_A1_to_B1 0.5 0 to 1 No k_B1 0.1 0 to (unbounded) No
Estimated Values: Parameter Value σ Prob. > t Lower
(90%) CI Upper
(90%) CI Lower
(95%) CI Upper (95%)
CI Parent_0 786.2 16.73 N/A 757.7 814.7 751.9 820.5 k_Parent 0.001566 6.30E-005 1.91E-020 0.001459 0.001673 0.001437 0.002 f_Parent_to_A1 0.8393 0.02498 N/A 0.7967 0.8818 0.788 0.891 k_A1 1.13E-004 4.71E-005 0.01196 3.25E-005 1.93E-004 1.61E-005 0 f_A1_to_B1 1.61E-015 16.33 N/A -27.81 27.81 -33.5 33.5 k_B1 0 0.008901 0.5 -0.01516 0.01516 -0.01826 0.018 f_Parent_to_B1 0.161 0.02491 N/A 0.1186 0.2034 0.1099 0.212
χ≤ Parameter Error % Degrees of Freedom
All data 6.87 27 Parent 4.38 10 A1 6.16 9 B1 14.7 8
Decay Times: Compartment DT50 (Minutes) DT90 (Minutes)
Parent 443 1.47E+03 A1 6.15E+03 >10,000 B1 >10,000 >10,000
0 200 400 600 800 1000 1200 14000
100
200
300
400
500
600
700
800
900
1000
Time (min)
µM
CPF CPF Oxon CPF Sum
0
200
400
600
800
Con
cent
rati
on (
nM)
0 200 400 600 800 1000 1200 1400
Time (Minutes)
Parent Parent Fit A1 A1 Fit B1 B1 Fit
0 200 400 600 800 1000 1200 14000
100
200
300
400
500
600
700
800
900
1000
Time (min)
µM
CPF CPF Oxon CPF Sum
0 500 1000 15000
100
200
300
400
500
600
700
800
900
1000
Time (min)
µM
CPF t1/2 = 443 minCPF Oxon t1/2 = 6150 minTCP t1/2 > 10000 min
Figure 62: CAKE software fit of CPF, CPF Oxon, and TCP concentration versus time. Calculated half-lives are presented in the inset.
190
The following data are reported from the CAKE software package using single
first order (SFO) kinetic fits of experimental CPF, CPF Oxon and TCP data.
Table 11: Initial values of CAKE sequence parameters
Parameter Initial Value Bounds Fixed Parent_0 781 0 to (unbounded) No k_Parent 0.1 0 to (unbounded) No f_Parent_to_A1 0.3333 0 to 1 No f_Parent_to_B1 0.3333 0 to 1 No k_A1 0.1 0 to (unbounded) No f_A1_to_B1 0.5 0 to 1 No k_B1 0.1 0 to (unbounded) No
Table 12: Estimated values from CAKE software.
Parameter Value s Prob. > t Lower (90%) CI
Upper (90%) CI
Parent_0 786.2 16.73 N/A 757.7 814.7 k_Parent 0.001566 6.30E-005 1.91E-020 0.001459 0.001673 f_Parent_to_A1 0.8393 0.02498 N/A 0.7967 0.8818 k_A1 1.13E-004 4.71E-005 0.01196 3.25E-005 1.93E-004 f_A1_to_B1 1.61E-015 16.33 N/A -27.81 27.81 k_B1 0 0.008901 0.5 -0.01516 0.01516 f_Parent_to_B1 0.161 0.02491 N/A 0.1186 0.2034
Table 13: χ ≤ the listed values from CAKE software.
Parameter Error % Degrees of Freedom All data 6.87 27 Parent 4.38 10 A1 6.16 9 B1 14.7 8
191
Table 14: Decay time from CAKE software.
Compartment DT50 (Minutes) DT90 (Minutes) Parent 443 1.47E+03 A1 6.15E+03 >10,000 B1 >10,000 >10,000
Table 15: Additional statistics from CAKE software.
Parameter r≤ (Obs v Pred) Efficiency All data 0.9954 0.995 Parent 0.9972 0.9776 A1 0.9964 0.995 B1 0.9757 0.973
Inactivation due to Degradation Products
No inactivation was seen following up to 60 minutes incubation in 1 µM CPF
Oxon or TCP. These concentrations were selected to test inactivation should complete
transformation of CPF into either of the degradation products occur, and all samples
were performed under dark conditions. The lack of inactivation indicates that neither of
the degradation products are acutely toxic to bacteria.
192
Additional Statistics: Parameter r≤ (Obs v Pred) Efficiency
All data 0.9954 0.995 Parent 0.9972 0.9776 A1 0.9964 0.995 B1 0.9757 0.973
6. Inactivation due to Degradation Products No inactivation was seen following up to 60 minutes incubation in 1µM CPF Oxon or TCP. These concentrations were selected to test inactivation should complete transformation of CPF into either of the degradation products occur, and all samples were performed under dark conditions. The lack of inactivation indicates that neither of the degradation products are acutely toxic to bacteria.
(a) (b)
Figure SI-6 – Bacterial inactivation tests under dark conditions in the presence of CPF Oxon and TCP. a) B. subtilis and b) A. baumanii.
0 10 20 30 40 50 600
0.2
0.4
0.6
0.8
1
1.2
1.4B.subtilis Inactivation
Time (min)
N/N
0
CPF Oxon TCP Control
0 10 20 30 40 50 600
0.2
0.4
0.6
0.8
1
1.2
1.4A.baumannii Inactivation
Time (min)
N/N
0
CPF Oxon TCP Control
Figure 63: Bacterial inactivation tests under dark conditions in the presence of CPF Oxon and TCP (a) B. subtilis and (b) A. baumanii.
193
References [1] M.F. Hochella, S.K. Lower, P.A. Maurice, R.L. Penn, N. Sahai, D.L. Sparks, B.S. Twining, Nanominerals, mineral nanoparticles, and earth systems, Science, 319 (2008) 1631-1635.
[2] Z. Su, W. Zhou, Y. Zhang, New insight into the soot nanoparticles in a candle flame, Chem Commun, 47 (2011) 4700-4702.
[3] J.L. Gardea-Torresdey, E. Gomez, J.R. Peralta-Videa, J.G. Parsons, H. Troiani, M. Jose-Yacaman, Alfalfa sprouts: a natural source for the synthesis of silver nanoparticles, Langmuir, 19 (2003) 1357-1361.
[4] N.F. Adegboyega, V.K. Sharma, K. Siskova, R. Zboril, M. Sohn, B.J. Schultz, S. Banerjee, Interactions of aqueous Ag+ with fulvic acids: mechanisms of silver nanoparticle formation and investigation of stability, Environ Sci Technol, 47 (2013) 757-764.
[5] A.R. Badireddy, J.F. Budarz, S.M. Marinakos, S. Chellam, M.R. Wiesner, Formation of Silver Nanoparticles in Visible Light-Illuminated Waters: Mechanism and Possible Impacts on the Persistence of AgNPs and Bacterial Lysis, Environ Eng Sci, 31 (2014) 338-349.
[6] F. Kang, P.J. Alvarez, D. Zhu, Microbial extracellular polymeric substances reduce Ag+ to silver nanoparticles and antagonize bactericidal activity, Environ Sci Technol, 48 (2014) 316-322.
[7] M.C. Roco, The long view of nanotechnology development: the National Nanotechnology Initiative at 10 years, J Nanopart Res, 13 (2011) 427-445.
[8] C.O. Hendren, X. Mesnard, J. Droge, M.R. Wiesner, Estimating production data for five engineered nanomaterials as a basis for exposure assessment, Environ Sci Technol, 45 (2011) 2562-2569.
[9] A.A. Keller, A. Lazareva, Predicted Releases of Engineered Nanomaterials: From Global to Regional to Local, Environmental Science & Technology Letters, 1 (2014) 65-70.
[10] J.-Y. Bottero, M. Auffan, D. Borschnek, P. Chaurand, J. Labille, C. Levard, A. Masion, M. Tella, J. Rose, M.R. Wiesner, Nanotechnology, global development in the frame of environmental risk forecasting. A necessity of interdisciplinary researches, Comptes Rendus Geoscience, 347 (2015) 35-42.
[11] A.A. Keller, S. McFerran, A. Lazareva, S. Suh, Global life cycle releases of engineered nanomaterials, J Nanopart Res, 15 (2013) 1-17.
[12] M.E. Vance, T. Kuiken, E.P. Vejerano, S.P. McGinnis, M.F. Hochella Jr, D. Rejeski, M.S. Hull, Nanotechnology in the real world: Redeveloping the nanomaterial consumer products inventory, Beilstein journal of nanotechnology, 6 (2015) 1769-1780.
194
[13] A. Fujishima, X. Zhang, D. Tryk, TiO2 photocatalysis and related surface phenomena, Surf Sci Rep, 63 (2008) 515-582.
[14] M.N. Chong, B. Jin, C.W. Chow, C. Saint, Recent developments in photocatalytic water treatment technology: a review, Water Res, 44 (2010) 2997-3027.
[15] M.R. Hoffmann, S.T. Martin, W.Y. Choi, D.W. Bahnemann, Environmental Applications of Semiconductor Photocatalysis, Chem Rev, 95 (1995) 69-96.
[16] M. Hoffmann, E.M. Hotze, M.R. Wiesner, Reactive oxygen species generation on nanoparticulate material, Environmental Nanotechnology. Applications and Impacts of Nanomaterials, (2007) 155-203.
[17] C.O. Robichaud, D. Tanzil, U. Weilenmann, M.R. Wiesner, Relative risk analysis of several manufactured nanomaterials: An insurance industry context, Environ Sci Technol, 39 (2005) 8985-8994.
[18] M.R. Wiesner, G.V. Lowry, P. Alvarez, D. Dionysiou, P. Biswas, Assessing the risks of manufactured nanomaterials, Environ Sci Technol, 40 (2006) 4336-4345.
[19] D. Hristozov, I. Malsch, Hazards and Risks of Engineered Nanoparticles for the Environment and Human Health, Sustainability, 1 (2009) 1161-1194.
[20] N.C. Mueller, B. Nowack, Exposure modeling of engineered nanoparticles in the environment, Environ Sci Technol, 42 (2008) 4447-4453.
[21] B. Nowack, T.D. Bucheli, Occurrence, behavior and effects of nanoparticles in the environment, Environmental pollution, 150 (2007) 5-22.
[22] J.A. Brant, J. Labille, C.O. Robichaud, M. Wiesner, Fullerol cluster formation in aqueous solutions: Implications for environmental release, J Colloid Interf Sci, 314 (2007) 281-288.
[23] W.-C. Hou, C.T. Jafvert, Photochemical transformation of aqueous C60 clusters in sunlight, Environ Sci Technol, 43 (2008) 362-367.
[24] J. Lee, M. Cho, J.D. Fortner, J.B. Hughes, J.H. Kim, Transformation of Aggregate C-60 in the Aqueous Phase by UV Irradiation, Environ Sci Technol, 43 (2009) 4878-4883.
[25] R.L. Johnson, J.T. Nurmi, G.S. O'Brien Johnson, D. Fan, R.L. O'Brien Johnson, Z. Shi, A.J. Salter-Blanc, P.G. Tratnyek, G.V. Lowry, Field-scale transport and transformation of carboxymethylcellulose-stabilized nano zero-valent iron, Environ Sci Technol, 47 (2013) 1573-1580.
[26] K.L. Chen, M. Elimelech, Aggregation and deposition kinetics of fullerene (C-60) nanoparticles, Langmuir, 22 (2006) 10994-11001.
195
[27] X.L. Qu, P.J.J. Alvarez, Q.L. Li, Impact of Sunlight and Humic Acid on the Deposition Kinetics of Aqueous Fullerene Nanoparticles (nC(60)), Environ Sci Technol, 46 (2012) 13455-13462.
[28] S.R. Chae, Y. Xiao, S. Lin, T. Noeiaghaei, J.O. Kim, M.R. Wiesner, Effects of humic acid and electrolytes on photocatalytic reactivity and transport of carbon nanoparticle aggregates in water, Water Res, 46 (2012) 4053-4062.
[29] E.M. Hotze, T. Phenrat, G.V. Lowry, Nanoparticle Aggregation: Challenges to Understanding Transport and Reactivity in the Environment, J Environ Qual, 39 (2010) 1909-1924.
[30] R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, Y. Taga, Visible-light photocatalysis in nitrogen-doped titanium oxides, Science, 293 (2001) 269-271.
[31] R. Daghrir, P. Drogui, D. Robert, Modified TiO2 for environmental photocatalytic applications: a review, Industrial & Engineering Chemistry Research, 52 (2013) 3581-3599.
[32] S.D. Snow, J. Lee, J.H. Kim, Photochemical and Photophysical Properties of Sequentially Functionalized Fullerenes in the Aqueous Phase, Environ Sci Technol, 46 (2012) 13227-13234.
[33] M. Cho, J. Lee, Y. Mackeyev, L.J. Wilson, P.J.J. Alvarez, J.B. Hughes, J.-H. Kim, Visible Light Sensitized Inactivation of MS-2 Bacteriophage by a Cationic Amine-Functionalized C60 Derivative, Environ. Sci. Technol., 44 (2010) 6685-6691.
[34] K.A. Brausch, T.A. Anderson, P.N. Smith, J.D. Maul, Effects of functionalized fullerenes on bifenthrin and tribufos toxicity to Daphnia magna: Survival, reproduction, and growth rate, Environ Toxicol Chem, 29 (2010) 2600-2606.
[35] M. Heinlaan, A. Ivask, I. Blinova, H.-C. Dubourguier, A. Kahru, Toxicity of nanosized and bulk ZnO, CuO and TiO 2 to bacteria Vibrio fischeri and crustaceans Daphnia magna and Thamnocephalus platyurus, Chemosphere, 71 (2008) 1308-1316.
[36] A. Fujishima, K. Honda, Electrochemical evidence for the mechanism of the primary stage of photosynthesis, Bulletin of the chemical society of Japan, 44 (1971) 1148-1150.
[37] K. Hashimoto, H. Irie, A. Fujishima, TiO2Photocatalysis: A Historical Overview and Future Prospects, Japanese Journal of Applied Physics, 44 (2005) 8269-8285.
[38] J.M. Coronado, A Historical Introduction to Photocatalysis, in: Design of Advanced Photocatalytic Materials for Energy and Environmental Applications, Springer, 2013, pp. 1-4.
[39] M.A. Henderson, A surface science perspective on photocatalysis, Surf Sci Rep, 66 (2011) 185-297.
[40] A.L. Linsebigler, G.Q. Lu, J.T. Yates, Photocatalysis on Tio2 Surfaces - Principles, Mechanisms, and Selected Results, Chem Rev, 95 (1995) 735-758.
196
[41] M. Pelaez, N.T. Nolan, S.C. Pillai, M.K. Seery, P. Falaras, A.G. Kontos, P.S. Dunlop, J.W. Hamilton, J.A. Byrne, K. O'shea, A review on the visible light active titanium dioxide photocatalysts for environmental applications, Applied Catalysis B: Environmental, 125 (2012) 331-349.
[42] J.-M. Herrmann, Heterogeneous photocatalysis: fundamentals and applications to the removal of various types of aqueous pollutants, Catalysis today, 53 (1999) 115-129.
[43] A.O. Ibhadon, P. Fitzpatrick, Heterogeneous photocatalysis: recent advances and applications, Catalysts, 3 (2013) 189-218.
[44] K. Nakata, A. Fujishima, TiO 2 photocatalysis: design and applications, Journal of Photochemistry and Photobiology C: Photochemistry Reviews, 13 (2012) 169-189.
[45] K. Nakata, T. Ochiai, T. Murakami, A. Fujishima, Photoenergy conversion with TiO 2 photocatalysis: new materials and recent applications, Electrochimica Acta, 84 (2012) 103-111.
[46] A. Kubacka, M. Fernandez-Garcia, G. Colon, Advanced nanoarchitectures for solar photocatalytic applications, Chem Rev, 112 (2011) 1555-1614.
[47] T.E. Agustina, H. Ang, V. Vareek, A review of synergistic effect of photocatalysis and ozonation on wastewater treatment, Journal of Photochemistry and Photobiology C: Photochemistry Reviews, 6 (2005) 264-273.
[48] Y. Lan, Y. Lu, Z. Ren, Mini review on photocatalysis of titanium dioxide nanoparticles and their solar applications, Nano Energy, 2 (2013) 1031-1045.
[49] R.P. Schwarzenbach, P.M. Gschwend, D.M. Imboden, Environmental organic chemistry, John Wiley & Sons, 2005.
[50] B. Singh, N. Sharma, Mechanistic implications of plastic degradation, Polymer Degradation and Stability, 93 (2008) 561-584.
[51] O.C. Zafiriou, J. Joussot-Dubien, R.G. Zepp, R.G. Zika, Photochemistry of natural waters, Environ Sci Technol, 18 (1984) 358A-371A.
[52] K.R. Wigginton, B.M. Pecson, T. Sigstam, F. Bosshard, T. Kohn, Virus inactivation mechanisms: impact of disinfectants on virus function and structural integrity, Environ Sci Technol, 46 (2012) 12069-12078.
[53] W. Hijnen, E. Beerendonk, G.J. Medema, Inactivation credit of UV radiation for viruses, bacteria and protozoan (oo) cysts in water: a review, Water Res, 40 (2006) 3-22.
[54] C. Wong, W. Chu, The direct photolysis and photocatalytic degradation of alachlor at different TiO 2 and UV sources, Chemosphere, 50 (2003) 981-987.
197
[55] V.J. Pereira, H.S. Weinberg, K.G. Linden, P.C. Singer, UV degradation kinetics and modeling of pharmaceutical compounds in laboratory grade and surface water via direct and indirect photolysis at 254 nm, Environ Sci Technol, 41 (2007) 1682-1688.
[56] G. Ruppert, R. Bauer, G. Heisler, The photo-Fenton reaction—an effective photochemical wastewater treatment process, Journal of Photochemistry and Photobiology A: Chemistry, 73 (1993) 75-78.
[57] R.G. Zepp, B.C. Faust, J. Hoigne, Hydroxyl Radical Formation in Aqueous Reactions (Ph 3-8) of Iron(Ii) with Hydrogen-Peroxide - the Photo-Fenton Reaction, Environ Sci Technol, 26 (1992) 313-319.
[58] N. Serpone, A. Salinaro, Terminology, relative photonic efficiencies and quantum yields in heterogeneous photocatalysis. Part I: Suggested protocol, Pure Appl Chem, 71 (1999) 303-320.
[59] F. Wilkinson, W.P. Helman, A.B. Ross, Quantum Yields for the Photosensitized Formation of the Lowest Electronically Excited Singlet-State of Molecular-Oxygen in Solution, J Phys Chem Ref Data, 22 (1993) 113-262.
[60] R.T. Di Giulio, P.C. Washburn, R.J. Wenning, G.W. Winston, C.S. Jewell, Biochemical responses in aquatic animals: a review of determinants of oxidative stress, Environ Toxicol Chem, 8 (1989) 1103-1123.
[61] E. Cabiscol, J. Tamarit, J. Ros, Oxidative stress in bacteria and protein damage by reactive oxygen species, International Microbiology, 3 (2010) 3-8.
[62] F.C. Fang, Antimicrobial reactive oxygen and nitrogen species: concepts and controversies, Nature reviews. Microbiology, 2 (2004) 820-832.
[63] C. Bogdan, M. Röllinghoff, A. Diefenbach, Reactive oxygen and reactive nitrogen intermediates in innate and specific immunity, Current opinion in immunology, 12 (2000) 64-76.
[64] G.W. Winston, R.T. Di Giulio, Prooxidant and antioxidant mechanisms in aquatic organisms, Aquatic toxicology, 19 (1991) 137-161.
[65] T. Kohn, M. Grandbois, K. McNeill, K.L. Nelson, Association with Natural Organic Matter Enhances the Sunlight-Mediated Inactivation of MS2 Coliphage by Singlet Oxygen, Environ Sci Technol, 41 (2007) 4626-4632.
[66] T. Kohn, K.L. Nelson, Sunlight-Mediated Inactivation of MS2 Coliphage via Exogenous Singlet Oxygen Produced by Sensitizers in Natural Waters, Environ Sci Technol, 41 (2007) 192-197.
[67] D.E. Latch, Microheterogenity of singlet oxygen distributions in irradiated humic acid solutions, Science, 311 (2006) 1743.
198
[68] S. Garg, A.L. Rose, T.D. Waite, Photochemical production of superoxide and hydrogen peroxide from natural organic matter, Geochim Cosmochim Ac, 75 (2011) 4310-4320.
[69] D.M. Leech, M.T. Snyder, R.G. Wetzel, Natural organic matter and sunlight accelerate the degradation of 17ß-estradiol in water, Sci Total Environ, 407 (2009) 2087-2092.
[70] G.V. Buxton, C.L. Greenstock, W.P. Helman, A.B. Ross, Critical-Review of Rate Constants for Reactions of Hydrated Electrons, Hydrogen-Atoms and Hydroxyl Radicals (.Oh/.O-) in Aqueous-Solution, J Phys Chem Ref Data, 17 (1988) 513-886.
[71] M. Cho, H. Chung, W. Choi, J. Yoon, Different inactivation behaviors of MS-2 phage and Escherichia coli in TiO2 photocatalytic disinfection, Applied and environmental microbiology, 71 (2005) 270-275.
[72] M.C. DeRosa, R.J. Crutchley, Photosensitized singlet oxygen and its applications, Coordin Chem Rev, 233 (2002) 351-371.
[73] G. Bartosz, Use of spectroscopic probes for detection of reactive oxygen species, Clinica chimica acta; international journal of clinical chemistry, 368 (2006) 53-76.
[74] T. Nagano, Bioimaging Probes for Reactive Oxygen Species and Reactive Nitrogen Species, J Clin Biochem Nutr, 45 (2009) 111-124.
[75] N. Soh, Recent advances in fluorescent probes for the detection of reactive oxygen species, Analytical and bioanalytical chemistry, 386 (2006) 532-543.
[76] K.-I. Ishibashi, A. Fujishima, T. Watanabe, K. Hashimoto, Detection of active oxidative species in TiO 2 photocatalysis using the fluorescence technique, Electrochemistry Communications, 2 (2000) 207-210.
[77] S.R. Chae, E.M. Hotze, M.R. Wiesner, Evaluation of the Oxidation of Organic Compounds by Aqueous Suspensions of Photosensitized Hydroxylated-C-60 Fullerene Aggregates, Environ Sci Technol, 43 (2009) 6208-6213.
[78] E.M. Hotze, A.R. Badireddy, S. Chellam, M.R. Wiesner, Mechanisms of Bacteriophage Inactivation via Singlet Oxygen Generation in UV Illuminated Fullerol Suspensions, Environ Sci Technol, 43 (2009) 6639-6645.
[79] Z. Wang, W. Ma, C. Chen, H. Ji, J. Zhao, Probing paramagnetic species in titania-based heterogeneous photocatalysis by electron spin resonance (ESR) spectroscopy—A mini review, Chemical Engineering Journal, 170 (2011) 353-362.
[80] D.M. Guldi, K.D. Asmus, Photophysical properties of mono- and multiply-functionalized fullerene derivatives, J Phys Chem A, 101 (1997) 1472-1481.
199
[81] L. Brunet, D. Lyon, E.M. Hotze, P.J.J. Alvarez, M.R. Wiesner, Comparative Photoactivity and Antibacterial Properties of C60 Fullerenes and Titanium Dioxide Nanoparticles, Environmental Science and Technology, 43 (2009) 4355-4360.
[82] S.R. Chae, M. Therezien, J.F. Budarz, L. Wessel, S.H. Lin, Y. Xiao, M.R. Wiesner, Comparison of the photosensitivity and bacterial toxicity of spherical and tubular fullerenes of variable aggregate size, J Nanopart Res, 13 (2011) 5121-5127.
[83] A. Gollmer, J. Arnbjerg, F.H. Blaikie, B.W. Pedersen, T. Breitenbach, K. Daasbjerg, M. Glasius, P.R. Ogilby, Singlet Oxygen Sensor Green(R): photochemical behavior in solution and in a mammalian cell, Photochemistry and photobiology, 87 (2011) 671-679.
[84] X. Ragas, A. Jimenez-Banzo, D. Sanchez-Garcia, X. Batllori, S. Nonell, Singlet oxygen photosensitisation by the fluorescent probe Singlet Oxygen Sensor Green, Chem Commun (Camb), (2009) 2920-2922.
[85] A. Okado-Matsumoto, I. Fridovich, Subcellular distribution of superoxide dismutases (SOD) in rat liver Cu, Zn-SOD in mitochondria, Journal of Biological Chemistry, 276 (2001) 38388-38393.
[86] H. Zhao, S. Kalivendi, H. Zhang, J. Joseph, K. Nithipatikom, J. Vásquez-Vivar, B. Kalyanaraman, Superoxide reacts with hydroethidine but forms a fluorescent product that is distinctly different from ethidium: potential implications in intracellular fluorescence detection of superoxide, Free Radical Biology and Medicine, 34 (2003) 1359-1368.
[87] K.-i. Ishibashi, A. Fujishima, T. Watanabe, K. Hashimoto, Quantum yields of active oxidative species formed on TiO 2 photocatalyst, Journal of Photochemistry and Photobiology A: Chemistry, 134 (2000) 139-142.
[88] S.E. Page, W.A. Arnold, K. McNeill, Terephthalate as a probe for photochemically generated hydroxyl radical, Journal of environmental monitoring : JEM, 12 (2010) 1658-1665.
[89] S. Eremia, D. Chevalierlucia, G. Radu, J. Marty, Optimization of hydroxyl radical formation using TiO2 as photocatalyst by response surface methodology, Talanta, 77 (2008) 858-862.
[90] C.S. Foote, R.W. Denny, Chemistry of Singlet Oxygen .7. Quenching by Beta-Carotene, J Am Chem Soc, 90 (1968) 6233-&.
[91] A.M. Wade, H.N. Tucker, Antioxidant characteristics of L-histidine, J Nutr Biochem, 9 (1998) 308-315.
[92] I.B.C. Matheson, J. Lee, Chemical-Reaction Rates of Amino-Acids with Singlet Oxygen, Photochemistry and photobiology, 29 (1979) 879-881.
[93] O.I. Aruoma, B. Halliwell, B.M. Hoey, J. Butler, The antioxidant action of N-acetylcysteine: its reaction with hydrogen peroxide, hydroxyl radical, superoxide, and hypochlorous acid, Free Radical Biology and Medicine, 6 (1989) 593-597.
200
[94] M. Gutiérrez, A. Henglein, F. Ibanez, Radical scavenging in the sonolysis of aqueous solutions of iodide, bromide, and azide, The Journal of Physical Chemistry, 95 (1991) 6044-6047.
[95] Y. Zhang, K. Aslan, M.J. Previte, C.D. Geddes, Plasmonic engineering of singlet oxygen generation, Proceedings of the National Academy of Sciences of the United States of America, 105 (2008) 1798-1802.
[96] D.C. Neckers, Rose-Bengal, J Photoch Photobio A, 47 (1989) 1-29.
[97] I.E. Kochevar, R.W. Redmond, Photosensitized production of singlet oxygen, Method Enzymol, 319 (2000) 20-28.
[98] H. Ukeda, S. Maeda, T. Ishii, M. Sawamura, Spectrophotometric assay for superoxide dismutase based on tetrazolium salt 3'-{1-[(phenylamino)-carbonyl]-3,4-tetrazolium}-bis(4-methoxy-6-nitro)benzenesulfonic acid hydrate reduction by xanthine-xanthine oxidase, Anal Biochem, 251 (1997) 206-209.
[99] W. Song, V. Ravindran, M. Pirbazari, Process optimization using a kinetic model for the ultraviolet radiation-hydrogen peroxide decomposition of natural and synthetic organic compounds in groundwater, Chemical Engineering Science, 63 (2008) 3249-3270.
[100] R. Zellner, M. Exner, H. Herrmann, Absolute Oh Quantum Yields in the Laser Photolysis of Nitrate, Nitrite and Dissolved H2o2 at 308 and 351 Nm in the Temperature-Range 278-353 K, J Atmos Chem, 10 (1990) 411-425.
[101] C. Flors, M.J. Fryer, J. Waring, B. Reeder, U. Bechtold, P.M. Mullineaux, S. Nonell, M.T. Wilson, N.R. Baker, Imaging the production of singlet oxygen in vivo using a new fluorescent sensor, Singlet Oxygen Sensor Green, Journal of experimental botany, 57 (2006) 1725-1734.
[102] F. Yan, Y. Zhang, K.S. Kim, H.K. Yuan, T. Vo-Dinh, Cellular Uptake and Photodynamic Activity of Protein Nanocages Containing Methylene Blue Photosensitizing Drug, Photochemistry and photobiology, 86 (2010) 662-666.
[103] P. Bilski, R. Dabestani, C.F. Chignell, Influence of Cationic Surfactant on the Photoprocesses of Eosin and Rose-Bengal in Aqueous-Solution, J Phys Chem-Us, 95 (1991) 5784-5791.
[104] C.L. Wu, K.G. Linden, Phototransformation of selected organophosphorus pesticides: Roles of hydroxyl and carbonate radicals, Water Res, 44 (2010) 3585-3594.
[105] C. Pierlot, V. Nardello, R. Schmidt, J.M. Aubry, Determination of physical (k(q)) and chemical (k(r)) rate constants for singlet oxygen quenching using the thermolysis of a naphthalenic endoperoxide in H2O and D2O, Arkivoc, (2007) 245-256.
[106] M. Benrahmoune, P. Thérond, Z. Abedinzadeh, The reaction of superoxide radical with N-acetylcysteine, Free Radical Biology and Medicine, 29 (2000) 775-782.
201
[107] M. Auffan, J. Rose, C. Chanéac, J.-P. Jolivet, A. Masion, M.R. Wiesner, J.-Y. Bottero, Surface reactivity of manufactured nanoparticles, in: Nanoethics and Nanotoxicology, Springer, 2011, pp. 269-290.
[108] D. Beydoun, R. Amal, G. Low, S. McEvoy, Role of nanoparticles in photocatalysis, J Nanopart Res, 1 (1999) 439-458.
[109] A. Sclafani, J.M. Herrmann, Influence of metallic silver and of platinum-silver bimetallic deposits on the photocatalytic activity of titania (anatase and rutile) in organic and aqueous media, J Photoch Photobio A, 113 (1998) 181-188.
[110] S. Banerjee, S.C. Pillai, P. Falaras, K.E. O’shea, J.A. Byrne, D.D. Dionysiou, New insights into the mechanism of visible light photocatalysis, The journal of physical chemistry letters, 5 (2014) 2543-2554.
[111] D.M. Guldi, M. Prato, Excited-state properties of C-60 fullerene derivatives, Accounts Chem Res, 33 (2000) 695-703.
[112] M.D. Diener, J.M. Alford, Isolation and properties of small-bandgap fullerenes, Nature, 393 (1998) 668-671.
[113] H.W. Kroto, A. Allaf, S. Balm, C60: Buckminsterfullerene, Chem Rev, 91 (1991) 1213-1235.
[114] D.J. Klein, T.G. Schmalz, G. Hite, W. Seitz, Resonance in C60 buckminsterfullerene, J Am Chem Soc, 108 (1986) 1301-1302.
[115] S. Leach, M. Vervloet, A. Després, E. Bréheret, J.P. Hare, T.J. Dennis, H.W. Kroto, R. Taylor, D.R. Walton, Electronic spectra and transitions of the fullerene C 60, Chemical Physics, 160 (1992) 451-466.
[116] J.A. Brant, J. Labille, J.Y. Bottero, M.R. Wiesner, Characterizing the impact of preparation method on fullerene cluster structure and chemistry, Langmuir, 22 (2006) 3878-3885.
[117] X.J. Chang, P.J. Vikesland, UV-vis Spectroscopic Properties of nC(60) Produced via Extended Mixing, Environ Sci Technol, 45 (2011) 9967-9974.
[118] D.Y. Lyon, L.K. Adams, J.C. Falkner, P.J. Alvarez, Antibacterial activity of fullerene water suspensions: effects of preparation method and particle size, Environ Sci Technol, 40 (2006) 4360-4366.
[119] D.Y. Lyon, L. Brunet, G.W. Hinkal, M.R. Wiesner, P.J. Alvarez, Antibacterial activity of fullerene water suspensions (nC60) is not due to ROS-mediated damage, Nano Lett, 8 (2008) 1539-1543.
202
[120] J. Brant, H. Lecoanet, M. Hotze, M. Wiesner, Comparison of electrokinetic properties of colloidal fullerenes (n-C60) formed using two procedures, Environ Sci Technol, 39 (2005) 6343-6351.
[121] S.R. Chae, A.R. Badireddy, J.F. Budarz, S.H. Lin, Y. Xiao, M. Therezien, M.R. Wiesner, Heterogeneities in Fullerene Nanoparticle Aggregates Affecting Reactivity, Bioactivity, and Transport, Acs Nano, 4 (2010) 5011-5018.
[122] J. Labille, A. Masion, F. Ziarelli, J. Rose, J. Brant, F. Villiéras, M. Pelletier, D. Borschneck, M.R. Wiesner, J.-Y. Bottero, Hydration and dispersion of C60 in aqueous systems: the nature of water− fullerene interactions, Langmuir, 25 (2009) 11232-11235.
[123] R. Taylor, D.R.M. Walton, The Chemistry of Fullerenes, Nature, 363 (1993) 685-693.
[124] V. Anantharaj, J. Bhonsle, T. Canteenwala, L.Y. Chiang, Highly electrophilic hexanitro[60]fullerenes, Elec Soc S, 98 (1998) 196-204.
[125] E. Nakamura, H. Isobe, Functionalized fullerenes in water. The first 10 years of their chemistry, biology, and nanoscience, Accounts Chem Res, 36 (2003) 807-815.
[126] T.H. Goswami, R. Singh, S. Alam, G.N. Mathur, Thermal analysis: a unique method to estimate the number of substituents in fullerene derivatives, Thermochim Acta, 419 (2004) 97-104.
[127] K.D. Pickering, M.R. Wiesner, Fullerol-sensitized production of reactive oxygen species in aqueous solution, Environ Sci Technol, 39 (2005) 1359-1365.
[128] A.R. Badireddy, E.M. Hotze, S. Chellam, P. Alvarez, M.R. Wiesner, Inactivation of Bacteriophages via photosensitization of fullerol nanoparticles, Environ Sci Technol, 41 (2007) 6627-6632.
[129] L.J. Kong, O. Tedrow, Y.F. Chan, R.G. Zepp, Light-Initiated Transformations of Fullerenol in Aqueous Media, Environ Sci Technol, 43 (2009) 9155-9160.
[130] J.W. Arbogast, A.P. Darmanyan, C.S. Foote, Y. Rubin, F.N. Diederich, M.M. Alvarez, S.J. Anz, R.L. Whetten, Photophysical Properties of C60, J Phys Chem-Us, 95 (1991) 11-12.
[131] C.S. Foote, Photophysical and Photochemical Properties of Fullerenes, Top Curr Chem, 169 (1994) 347-363.
[132] J. Lee, Y. Yamakoshi, J.B. Hughes, J.H. Kim, Mechanism of C-60 photoreactivity in water: Fate of triplet state and radical anion and production of reactive oxygen species, Environ Sci Technol, 42 (2008) 3459-3464.
[133] T. Hamano, K. Okuda, T. Mashino, M. Hirobe, K. Arakane, A. Ryu, S. Mashiko, T. Nagano, Singlet oxygen production from fullerene derivatives: Effect of sequential functionalization of the fullerene core, Chem Commun, (1997) 21-22.
203
[134] R. Bakry, R.M. Vallant, M. Najam-ul-Haq, M. Rainer, Z. Szabo, C.W. Huck, G.K. Bonn, Medicinal applications of fullerenes, International journal of nanomedicine, 2 (2007) 639.
[135] S. Bosi, T. Da Ros, G. Spalluto, M. Prato, Fullerene derivatives: an attractive tool for biological applications, European Journal of Medicinal Chemistry, 38 (2003) 913-923.
[136] S.Z. Wang, R.M. Gao, F.M. Zhou, M. Selke, Nanomaterials and singlet oxygen photosensitizers: potential applications in photodynamic therapy, J Mater Chem, 14 (2004) 487-493.
[137] J. Lee, Y. Mackeyev, M. Cho, D. Li, J.H. Kim, L.J. Wilson, P.J.J. Alvarez, Photochemical and Antimicrobial Properties of Novel C-60 Derivatives in Aqueous Systems, Environ Sci Technol, 43 (2009) 6604-6610.
[138] M. Cho, J. Lee, Y. Mackeyev, L.J. Wilson, P.J.J. Alvarez, J.B. Hughes, J.H. Kim, Visible Light Sensitized Inactivation of MS-2 Bacteriophage by a Cationic Amine-Functionalized C-60 Derivative, Environ Sci Technol, 44 (2010) 6685-6691.
[139] M. Kovochich, B. Espinasse, M. Auffan, E.M. Hotze, L. Wessel, T. Xia, A.E. Nel, M.R. Wiesner, Comparative Toxicity of C-60 Aggregates toward Mammalian Cells: Role of Tetrahydrofuran (THF) Decomposition, Environ Sci Technol, 43 (2009) 6378-6384.
[140] J. Lee, W.H. Song, S.S. Jang, J.D. Fortner, P.J.J. Alvarez, W.J. Cooper, J.H. Kim, Stability of Water-Stable C-60 Clusters to OH Radical Oxidation and Hydrated Electron Reduction, Environ Sci Technol, 44 (2010) 3786-3792.
[141] W.C. Hou, C.T. Jafvert, Photochemistry of Aqueous C-60 Clusters: Evidence of O-1(2) Formation and its Role in Mediating C-60 Phototransformation, Environ Sci Technol, 43 (2009) 5257-5262.
[142] Y.S. Hwang, Q.L. Li, Characterizing Photochemical Transformation of Aqueous nC(60) under Environmentally Relevant Conditions, Environ Sci Technol, 44 (2010) 3008-3013.
[143] U. Diebold, Structure and properties of TiO 2 surfaces: a brief review, Applied Physics A: Materials Science & Processing, 76 (2003) 681-687.
[144] C.J. Barbe, F. Arendse, P. Comte, M. Jirousek, F. Lenzmann, V. Shklover, M. Gratzel, Nanocrystalline titanium oxide electrodes for photovoltaic applications, J Am Ceram Soc, 80 (1997) 3157-3171.
[145] D.M. Blake, P.C. Maness, Z. Huang, E.J. Wolfrum, J. Huang, W.A. Jacoby, Application of the photocatalytic chemistry of titanium dioxide to disinfection and the killing of cancer cells, Separ Purif Method, 28 (1999) 1-50.
[146] M. Lazar, S. Varghese, S. Nair, Photocatalytic Water Treatment by Titanium Dioxide: Recent Updates, Catalysts, 2 (2012) 572-601.
204
[147] A. Fujishima, T.N. Rao, D.A. Tryk, Titanium dioxide photocatalysis, Journal of Photochemistry and Photobiology C: Photochemistry Reviews, 1 (2000) 1-21.
[148] T. Luttrell, S. Halpegamage, J. Tao, A. Kramer, E. Sutter, M. Batzill, Why is anatase a better photocatalyst than rutile?--Model studies on epitaxial TiO2 films, Scientific reports, 4 (2014) 4043.
[149] X. Kang, S. Chen, Photocatalytic reduction of methylene blue by TiO2 nanotube arrays: effects of TiO2 crystalline phase, Journal of Materials Science, 45 (2010) 2696-2702.
[150] H. Kominami, J.-i. Kato, S.-y. Murakami, Y. Ishii, M. Kohno, K.-i. Yabutani, T. Yamamoto, Y. Kera, M. Inoue, T. Inui, B. Ohtani, Solvothermal syntheses of semiconductor photocatalysts of ultra-high activities, Catalysis Today, 84 (2003) 181-189.
[151] B. Ohtani, J.-i. Handa, S.-i. Nishimoto, T. Kagiya, Highly active semiconductor photocatalyst: Extra-fine crystallite of brookite TiO 2 for redox reaction in aqueous propan-2-ol and/or silver sulfate solution, Chem Phys Lett, 120 (1985) 292-294.
[152] S. Bakardjieva, V. Stengl, L. Szatmary, J. Subrt, J. Lukac, N. Murafa, D. Niznansky, K. Cizek, J. Jirkovsky, N. Petrova, Transformation of brookite-type TiO2 nanocrystals to rutile: correlation between microstructure and photoactivity, J Mater Chem, 16 (2006) 1709.
[153] S. Bakardjieva, J. Šubrt, V. Štengl, M.J. Dianez, M.J. Sayagues, Photoactivity of anatase–rutile TiO2 nanocrystalline mixtures obtained by heat treatment of homogeneously precipitated anatase, Applied Catalysis B: Environmental, 58 (2005) 193-202.
[154] T. Hirakawa, Y. Nosaka, Properties of O(2)(center dot-) and OH(center dot) formed in TiO(2) aqueous suspensions by photocatalytic reaction and the influence of H(2)O(2) and some ions, Langmuir, 18 (2002) 3247-3254.
[155] D.C. Hurum, K.A. Gray, T. Rajh, M.C. Thurnauer, Recombination pathways in the Degussa P25 formulation of TiO2: Surface versus lattice mechanisms, J Phys Chem B, 109 (2005) 977-980.
[156] L.Z. Sun, J.R. Bolton, Determination of the quantum yield for the photochemical generation of hydroxyl radicals in TiO2 suspensions, J Phys Chem-Us, 100 (1996) 4127-4134.
[157] A. Turolla, M. Fumagalli, M. Bestetti, M. Antonelli, Electrophotocatalytic decolorization of an azo dye on TiO2 self-organized nanotubes in a laboratory scale reactor, Desalination, 285 (2012) 377-382.
[158] G. Mascolo, R. Ciannarella, L. Balest, A. Lopez, Effectiveness of UV-based advanced oxidation processes for the remediation of hydrocarbon pollution in the groundwater: a laboratory investigation, J Hazard Mater, 152 (2008) 1138-1145.
[159] C. Kormann, D.W. Bahnemann, M.R. Hoffmann, Photolysis of Chloroform and Other Organic-Molecules in Aqueous Tio2 Suspensions, Environ Sci Technol, 25 (1991) 494-500.
205
[160] S. Lakshmi, R. Renganathan, S. Fujita, Study on Tio2-Mediated Photocatalytic Degradation of Methylene-Blue, J Photoch Photobio A, 88 (1995) 163-167.
[161] V.K. Sharma, Aggregation and toxicity of titanium dioxide nanoparticles in aquatic environment—a review, Journal of Environmental Science and Health Part A, 44 (2009) 1485-1495.
[162] K. Sunada, Y. Kikuchi, K. Hashimoto, A. Fujishima, Bactericidal and detoxification effects of TiO2 thin film photocatalysts, Environ Sci Technol, 32 (1998) 726-728.
[163] Q. Li, S. Mahendra, D.Y. Lyon, L. Brunet, M.V. Liga, D. Li, P.J. Alvarez, Antimicrobial nanomaterials for water disinfection and microbial control: potential applications and implications, Water Res, 42 (2008) 4591-4602.
[164] D. Feng, N. Xie, C. Gong, Z. Leng, H. Xiao, H. Li, X. Shi, Portland cement paste modified by TiO2 nanoparticles: a microstructure perspective, Industrial & Engineering Chemistry Research, 52 (2013) 11575-11582.
[165] R. Khataee, V. Heydari, L. Moradkhannejhad, M. Safarpour, S. Joo, Self-cleaning and mechanical properties of modified white cement with nanostructured TiO2, J Nanosci Nanotechno, 13 (2013) 5109-5114.
[166] A. Maury-Ramirez, W. De Muynck, R. Stevens, K. Demeestere, N. De Belie, Titanium dioxide based strategies to prevent algal fouling on cementitious materials, Cement and Concrete Composites, 36 (2013) 93-100.
[167] A. Yousefi, A. Allahverdi, P. Hejazi, Effective dispersion of nano-TiO 2 powder for enhancement of photocatalytic properties in cement mixes, Construction and building materials, 41 (2013) 224-230.
[168] D. Sanchez-Quiles, A. Tovar-Sanchez, Sunscreens as a source of hydrogen peroxide production in coastal waters, Environ Sci Technol, 48 (2014) 9037-9042.
[169] A.P. Gondikas, F.v.d. Kammer, R.B. Reed, S. Wagner, J.F. Ranville, T. Hofmann, Release of TiO2 nanoparticles from sunscreens into surface waters: a one-year survey at the old Danube recreational Lake, Environ Sci Technol, 48 (2014) 5415-5422.
[170] A. Praetorius, J. Labille, M. Scheringer, A. Thill, K. Hungerbuhler, J.Y. Bottero, Heteroaggregation of Titanium Dioxide Nanoparticles with Model Natural Colloids under Environmentally Relevant Conditions, Environ Sci Technol, 48 (2014) 10690-10698.
[171] L. Windler, C. Lorenz, N. Von Goetz, K. Hungerbuhler, M. Amberg, M. Heuberger, B. Nowack, Release of titanium dioxide from textiles during washing, Environ Sci Technol, 46 (2012) 8181-8188.
206
[172] M. Fouqueray, P. Noury, L. Dherret, P. Chaurand, K. Abbaci, J. Labille, J. Rose, J. Garric, Exposure of juvenile Danio rerio to aged TiO2 nanomaterial from sunscreen, Environmental Science and Pollution Research, 20 (2013) 3340-3350.
[173] A. Al-Kattan, A. Wichser, S. Zuin, Y. Arroyo, L. Golanski, A. Ulrich, B. Nowack, Behavior of TiO2 released from nano-TiO2-containing paint and comparison to pristine nano-TiO2, Environ Sci Technol, 48 (2014) 6710-6718.
[174] C. Santaella, B. Allainmat, F. Simonet, C. Chanéac, J.r. Labille, M.l. Auffan, J.r. Rose, W. Achouak, Aged TiO2-based nanocomposite used in sunscreens produces singlet oxygen under long-wave UV and sensitizes Escherichia coli to cadmium, Environ Sci Technol, 48 (2014) 5245-5253.
[175] T. Ohno, K. Sarukawa, K. Tokieda, M. Matsumura, Morphology of a TiO2 Photocatalyst (Degussa, P-25) Consisting of Anatase and Rutile Crystalline Phases, J Catal, 203 (2001) 82-86.
[176] K.J.A. Raj, B. Viswanathan, Effect of surface area, pore volume and particle size of P25 titania on the phase transformation of anatase to rutile, Indian J Chem A, 48 (2009) 1378-1382.
[177] D. Jassby, J.F. Budarz, M. Wiesner, Impact of Aggregate Size and Structure on the Photocatalytic Properties of TiO2 and ZnO Nanoparticles, Environ Sci Technol, 46 (2012) 6934-6941.
[178] J. Taurozzi, V. Hackley, M. Wiesner, Preparation of a nanoscale TiO2 aqueous dispersion for toxicological or environmental testing, NIST Special Publication, 1200 (2012) 3.
[179] D. Thomas, S. Judd, N. Fawcett, Flocculation modelling: a review, Water Res, 33 (1999) 1579-1592.
[180] B. Derjaguin, L. Landau, The theory of stability of highly charged lyophobic sols and coalescence of highly charged particles in electrolyte solutions, Acta Physicochim. URSS, 14 (1941) 633-652.
[181] E. Verwey, J.T.G. Overbeek, Long distance forces acting between colloidal particles, Trans. Faraday Soc., 42 (1946) B117-B123.
[182] J.C. Crittenden, R.R. Trussell, D.W. Hand, K.J. Howe, G. Tchobanoglous, MWH's Water Treatment: Principles and Design: Principles and Design, John Wiley & Sons, 2012.
[183] K.L. Chen, S.E. Mylon, M. Elimelech, Aggregation kinetics of alginate-coated hematite nanoparticles in monovalent and divalent electrolytes, Environ Sci Technol, 40 (2006) 1516-1523.
[184] B. Espinasse, E.M. Hotze, M.R. Wiesner, Transport and retention of colloidal aggregates of C-60 in porous media: Effects of organic macromolecules, ionic composition, and preparation method, Environ Sci Technol, 41 (2007) 7396-7402.
207
[185] K.L. Chen, M. Elimelech, Interaction of Fullerene (C-60) Nanoparticles with Humic Acid and Alginate Coated Silica Surfaces: Measurements, Mechanisms, and Environmental Implications, Environ Sci Technol, 42 (2008) 7607-7614.
[186] Z.Q. Li, K. Greden, P.J.J. Alvarez, K.B. Gregory, G.V. Lowry, Adsorbed Polymer and NOM Limits Adhesion and Toxicity of Nano Scale Zerovalent Iron to E. coli, Environ Sci Technol, 44 (2010) 3462-3467.
[187] K.L. Chen, M. Elimelech, Influence of humic acid on the aggregation kinetics of fullerene (C-60) nanoparticles in monovalent and divalent electrolyte solutions, J Colloid Interf Sci, 309 (2007) 126-134.
[188] F. He, D.Y. Zhao, Preparation and characterization of a new class of starch-stabilized bimetallic nanoparticles for degradation of chlorinated hydrocarbons in water, Environ Sci Technol, 39 (2005) 3314-3320.
[189] S.H. Lin, Y.W. Cheng, J. Liu, M.R. Wiesner, Polymeric Coatings on Silver Nanoparticles Hinder Autoaggregation but Enhance Attachment to Uncoated Surfaces, Langmuir, 28 (2012) 4178-4186.
[190] K. Li, Y. Chen, Effect of natural organic matter on the aggregation kinetics of CeO 2 nanoparticles in KCl and CaCl 2 solutions: measurements and modeling, J Hazard Mater, 209 (2012) 264-270.
[191] C.V. Chrysikopoulos, V.I. Syngouna, Attachment of bacteriophages MS2 and ΦX174 onto kaolinite and montmorillonite: Extended-DLVO interactions, Colloids and Surfaces B: Biointerfaces, 92 (2012) 74-83.
[192] M.V. Smoluchowski, Smoluchowski, Versuch einer Mathematischen Theorie der Koagulationskinetic Kolloider Lösunger, Z. Phys. Chem, 92 (1917) 129-168.
[193] D.L. Swift, S.K. Friedlander, The coagulation of hydrosols by Brownian motion and laminar shear flow, Journal of colloid science, 19 (1964) 621-647.
[194] J.A. Raper, R. Amal, Measurement of Aggregate Fractal Dimensions Using Static Light-Scattering, Part Part Syst Char, 10 (1993) 239-245.
[195] M.Y. Lin, H.M. Lindsay, D.A. Weitz, R.C. Ball, R. Klein, P. Meakin, Universality in Colloid Aggregation, Nature, 339 (1989) 360-362.
[196] C.M. Sorensen, Light scattering by fractal aggregates: A review, Aerosol Sci Tech, 35 (2001) 648-687.
[197] S. Veerapaneni, M.R. Wiesner, Particle deposition on an infinitely permeable surface: Dependence of deposit morphology on particle size, J Colloid Interf Sci, 162 (1994) 110-122.
208
[198] J. Zhang, J. Buffle, Multi-method determination of the fractal dimension of hematite aggregates, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 107 (1996) 175-187.
[199] A. Turolla, A. Piazzoli, J. Farner Budarz, M.R. Wiesner, M. Antonelli, Experimental measurement and modelling of reactive species generation in TiO2 nanoparticle photocatalysis, Chemical Engineering Journal, 271 (2015) 260-268.
[200] Z. Meng, S.M. Hashmi, M. Elimelech, Aggregation rate and fractal dimension of fullerene nanoparticles via simultaneous multiangle static and dynamic light scattering measurement, J Colloid Interf Sci, 392 (2013) 27-33.
[201] T. Rudalevige, A.H. Francis, R. Zand, Spectroscopic studies of fullerene aggregates, J Phys Chem A, 102 (1998) 9797-9802.
[202] D. Zhou, A.A. Keller, Role of morphology in the aggregation kinetics of ZnO nanoparticles, Water Res, 44 (2010) 2948-2956.
[203] A.Y. Kim, J.C. Berg, Fractal heteroaggregation of oppositely charged colloids, J Colloid Interf Sci, 229 (2000) 607-614.
[204] E. Barbot, P. Dussouillez, J.-Y. Bottero, P. Moulin, Coagulation of bentonite suspension by polyelectrolytes or ferric chloride: Floc breakage and reformation, Chemical Engineering Journal, 156 (2010) 83-91.
[205] V. Oles, Shear-induced aggregation and breakup of polystyrene latex particles, J Colloid Interf Sci, 154 (1992) 351-358.
[206] P. Adler, Streamlines in and around porous particles, J Colloid Interf Sci, 81 (1981) 531-535.
[207] D.H. Li, J. Ganczarczyk, Fractal geometry of particle aggregates generated in water and wastewater treatment processes, Environ Sci Technol, 23 (1989) 1385-1389.
[208] S. Chellam, M.R. Wiesner, Fluid mechanics and fractal aggregates, Water Res, 27 (1993) 1493-1496.
[209] S. Veerapaneni, M.R. Wiesner, Hydrodynamics of fractal aggregates with radially varying permeability, J Colloid Interf Sci, 177 (1996) 45-57.
[210] M. Han, D.F. Lawler, The (relative) insignificance of G in flocculation, Journal (American Water Works Association), (1992) 79-91.
[211] C.P. Johnson, X. Li, B.E. Logan, Settling velocities of fractal aggregates, Environ Sci Technol, 30 (1996) 1911-1918.
209
[212] X. Li, B.E. Logan, Collision frequencies of fractal aggregates with small particles by differential sedimentation, Environ Sci Technol, 31 (1997) 1229-1236.
[213] K.H. Gardner, T.L. Theis, T.C. Young, The significance of shear stress in the agglomeration kinetics of fractal aggregates, Water Res, 32 (1998) 2660-2668.
[214] A.K. Yeung, R. Pelton, Micromechanics: a new approach to studying the strength and breakup of flocs, J Colloid Interf Sci, 184 (1996) 579-585.
[215] G. Bushell, Y. Yan, D. Woodfield, J. Raper, R. Amal, On techniques for the measurement of the mass fractal dimension of aggregates, Adv Colloid Interfac, 95 (2002) 1-50.
[216] B. Gilbert, R.K. Ono, K.A. Ching, C.S. Kim, The effects of nanoparticle aggregation processes on aggregate structure and metal uptake, J Colloid Interf Sci, 339 (2009) 285-295.
[217] D. Jassby, M. Wiesner, Characterization of ZnS nanoparticle aggregation using photoluminescence, Langmuir, 27 (2011) 902-908.
[218] J. Lee, J.D. Fortner, J.B. Hughes, J.H. Kim, Photochemical production of reactive oxygen species by C-60 in the aqueous phase during UV irradiation, Environ Sci Technol, 41 (2007) 2529-2535.
[219] E.M. Hotze, J.Y. Bottero, M.R. Wiesner, Theoretical Framework for Nanoparticle Reactivity as a Function of Aggregation State, Langmuir, 26 (2010) 11170-11175.
[220] F. Prat, R. Stackow, R. Bernstein, W.Y. Qian, Y. Rubin, C.S. Foote, Triplet-state properties and singlet oxygen generation in a homologous series of functionalized fullerene derivatives, J Phys Chem A, 103 (1999) 7230-7235.
[221] E.M. Hotze, J. Labille, P. Alvarez, M.R. Wiesner, Mechanisms of photochemistry and reactive oxygen production by fullerene suspensions in water, Environ Sci Technol, 42 (2008) 4175-4180.
[222] L. Brunet, D.Y. Lyon, E.M. Hotze, P.J.J. Alvarez, M.R. Wiesner, Comparative Photoactivity and Antibacterial Properties of C-60 Fullerenes and Titanium Dioxide Nanoparticles, Environ Sci Technol, 43 (2009) 4355-4360.
[223] M.J. Paul, J.L. Meyer, Streams in the urban landscape, Annual review of Ecology and Systematics, (2001) 333-365.
[224] A. Nikanorov, L. Brazhnikova, Water chemical composition of rives, lakes and wetlands, Types and Properties of Water, 2 (2009) 42-80.
[225] D.B. Medinas, G. Cerchiaro, D.F. Trindade, O. Augusto, The carbonate radical and related oxidants derived from bicarbonate buffer, IUBMB life, 59 (2007) 255-262.
210
[226] N.S. Wigginton, K.L. Haus, M.F. Hochella, Jr., Aquatic environmental nanoparticles, Journal of environmental monitoring : JEM, 9 (2007) 1306-1316.
[227] P. Connor, A.J. McQuillan, Phosphate adsorption onto TiO2 from aqueous solutions: an in situ internal reflection infrared spectroscopic study, Langmuir, 15 (1999) 2916-2921.
[228] R.F. Domingos, C. Peyrot, K.J. Wilkinson, Aggregation of titanium dioxide nanoparticles: role of calcium and phosphate, Environmental Chemistry, 7 (2010) 61-66.
[229] J.A. Leenheer, J.P. Croue, Characterizing aquatic dissolved organic matter, Environ Sci Technol, 37 (2003) 18A-26A.
[230] G. VanLoon, S. Duffy, Organic matter in water, Environmental Chemistry: A Global Perspective, 2nd edn., Oxford University Press, New York, (2005).
[231] R.F. Domingos, N. Tufenkji, K.J. Wilkinson, Aggregation of titanium dioxide nanoparticles: role of a fulvic acid, Environ Sci Technol, 43 (2009) 1282-1286.
[232] J.T. Quik, I. Velzeboer, M. Wouterse, A.A. Koelmans, D. van de Meent, Heteroaggregation and sedimentation rates for nanomaterials in natural waters, Water Res, 48 (2014) 269-279.
[233] J.T.K. Quik, M.C. Stuart, M. Wouterse, W. Peijnenburg, A.J. Hendriks, D. van de Meent, Natural colloids are the dominant factor in the sedimentation of nanoparticles, Environ Toxicol Chem, 31 (2012) 1019-1022.
[234] J.M. Lopez-Lopez, A. Schmitt, A. Moncho-Jorda, R. Hidalgo-Alvarez, Stability of binary colloids: kinetic and structural aspects of heteroaggregation processes, Soft Matter, 2 (2006) 1025-1042.
[235] P.D. Yates, G.V. Franks, S. Biggs, G.J. Jameson, Heteroaggregation with nanoparticles: effect of particle size ratio on optimum particle dose, Colloid Surface A, 255 (2005) 85-90.
[236] H.R. Jafry, M.V. Liga, Q. Li, A.R. Barron, Simple route to enhanced photocatalytic activity of p25 titanium dioxide nanoparticles by silica addition, Environ Sci Technol, 45 (2011) 1563-1568.
[237] M.V. Liga, S.J. Maguire-Boyle, H.R. Jafry, A.R. Barron, Q.L. Li, Silica Decorated TiO2 for Virus Inactivation in Drinking Water - Simple Synthesis Method and Mechanisms of Enhanced Inactivation Kinetics, Environ Sci Technol, 47 (2013) 6463-6470.
[238] N. Siedl, S.O. Baumann, M.J. Elser, O. Diwald, Particle Networks from Powder Mixtures: Generation of TiO2-SnO2 Heterojunctions via Surface Charge-Induced Heteroaggregation, J Phys Chem C, 116 (2012) 22967-22973.
[239] C.B. Almquist, P. Biswas, Role of Synthesis Method and Particle Size of Nanostructured TiO2 on Its Photoactivity, J Catal, 212 (2002) 145-156.
211
[240] A. Nel, T. Xia, L. Madler, N. Li, Toxic potential of materials at the nanolevel, Science, 311 (2006) 622-627.
[241] C.O. Hendren, C.M. Powers, M.D. Hoover, S.L. Harper, The Nanomaterial Data Curation Initiative: A collaborative approach to assessing, evaluating, and advancing the state of the field, Beilstein journal of nanotechnology, 6 (2015) 1752-1762.
[242] P.A. Holden, J. Gardea-Torresdey, F. Klaessig, R.F. Turco, M. Mortimer, K. Hund-Rinke, E.A. Cohen Hubal, D. Avery, D. Barcelo, R. Behra, Considerations of Environmentally Relevant Test Conditions for Improved Evaluation of Ecological Hazards of Engineered Nanomaterials, Environ Sci Technol, (2016).
[243] M.R. Wiesner, G.V. Lowry, K.L. Jones, J. Hochella, Michael F, R.T. Di Giulio, E. Casman, E.S. Bernhardt, Decreasing Uncertainties in assessing environmental exposure, risk, and ecological implications of nanomaterials†‡, Environ Sci Technol, 43 (2009) 6458-6462.
[244] C.O. Hendren, G.V. Lowry, J.M. Unrine, M.R. Wiesner, A functional assay-based strategy for nanomaterial risk forecasting, Sci Total Environ, 536 (2015) 1029-1037.
[245] G.V. Lowry, B.P. Espinasse, A.R. Badireddy, C.J. Richardson, B.C. Reinsch, L.D. Bryant, A.J. Bone, A. Deonarine, S. Chae, M. Therezien, B.P. Colman, H. Hsu-Kim, E.S. Bernhardt, C.W. Matson, M.R. Wiesner, Long-Term Transformation and Fate of Manufactured Ag Nanoparticles in a Simulated Large Scale Freshwater Emergent Wetland, Environ Sci Technol, 46 (2012) 7027-7036.
[246] A.R. Badireddy, E.M. Hotze, S. Chellam, P.J.J. Alveraz, M.R. Wiesner, Inactivation of bacteriophages via photosensitization of fullerol nanoparticles, Environ. Sci. Technol., 41 (2007) 6627-6632.
[247] E.M. Hotze, A.R. Badireddy, S. Chellam, M.R. Wiesner, Mechanisms of Bacteriophage Inactivation via Singlet Oxygen Generation in UV Illuminated Fullerol Suspensions, Environ. Sci. Technol., 43 (2009) 6639-6645.
[248] L. Huang, M. Terakawa, T. Zhiyentayev, Y.-Y. Huang, Y. Sawayama, A. Jahnke, G.P. Tegos, T. Wharton, M.R. Hamblin, Innovative Cationic Fullerenes as Broad-Spectrum Light-Activated Antimicrobials, Nanomedicine: Nanotechnology, Biology, and Medicine, 6 (2010) 442-452.
[249] P. Morz, G.P. Tegos, H. Gali, T. Wharton, T. Sarna, M.R. Hamblin, Photodynamic therapy with fullerenes, Photochem. Photobiol. Sci., 6 (2007) 1139-1149.
[250] M.R. Hamblin, G. Jori, Photodynamic Inactivation of Microbial Pathogens: Medical and Environmental Applications, The Royal Society of Chemistry, Cambridge, UK, 2011.
212
[251] T. Kohn, M. Grandbois, K. McNeill, K.L. Nelson, Association with natural organic matter enhances the sunlight-mediated inactivation of MS2 coliphage by singlet oxygen, Environ. Sci. Technol., 41 (2007) 4626-4632.
[252] X.M. Yan, B.Y. Shi, D.S. Wang, H.X. Tang, The Eco-toxic aspects of aqueous nano-C-60 Fullerenes, Progr. Chem., 20 (2008) 422-428.
[253] J. Lee, M. Cho, J.D. Fortner, J.B. Hughes, J.H. Kim, Transformation of Aggregate C(60) in the Aqueous Phase by UV Irradiation, Environmental Science & Technology, 43 (2009) 4878-4883.
[254] E.M. Hotze, J. Labille, P.J.J. Alvarez, M. Wiesner, Mechanisms of photochemistry and reactive oxygen production by fullerene suspensions in water, Environ. Sci. Technol., 42 (2008) 4175-4180.
[255] P. Thevenot, J. Cho, D. Wavhal, R.B. Timmons, L.P. Tang, Surface chemistry influences cancer killing effect of TiO(2) nanoparticles, Nanomedicine-Nanotechnology Biology and Medicine, 4 (2008) 226-236.
[256] M.A. Rodgers, P.T. Snowden, Lifetime of O2(1Δg) in Liquid Water As Determined by Time-Resolved Infrared Luminescence Measurements, J. Amer. Chem. Soc., 104 (1982) 5541-5543.
[257] R.W. Redmond, I.E. Kochevar, Spatially Resolved Cellular Responses to Singlet Oxygen, Photochem. Photobiol., 82 (2006) 1178-1186.
[258] M. Cho, H. Chung, W. Choi, J. Yoon, Different inactivation behaviors of MS-2 phage and Escherichia Coli in TiO2 photocatalytic disinfection, Appl. Environ. Microbiol., 71 (2005) 270-275.
[259] F. Rancan, S. Rosan, F. Boehm, A. Cantrell, M. Brellreich, H. Schoenberger, A. Hirsch, F. Moussa, Cytotoxicity and Photocytotoxicity of a Dendritic C60 mono-adduct and a Malonic Acid C60 Tris-adduct on Jurkat Cells, J. Photochem. Photobiol B: Biol., 67 (2002) 157-162.
[260] F. Käsermann, C. Kempf, Buckminsterfullerene and photodynamic inactivation of viruses, Rev. Med. Virol., 8 (1998) 143-151.
[261] J.Y. Kim, C. Lee, D.C. Love, D.L. Sedlak, J. Yoon, K.L. Nelson, Inactivation of MS2 Coliphage by Ferrous Ion and Zero-Valent Iron Nanoparticles, Environ. Sci. Technol., 45 (2011) 6978-6984.
[262] K.R. Wigginton, L. Menin, J.P. Montoya, T. Kohn, Oxidation of Virus Proteins during UV254 and Singlet Oxygen Mediated Inactivation, Environ. Sci. Technol., 44 (2010) 5437-5443.
[263] H. Grahn, P. Geladi, Techniques and Applications of Hyperspectral Image Analysis, John Wiley and Sons Ltd, Chichester, West Sussex, 2009.
[264] A. Barth, C. Zscherp, What vibrations tell us about proteins?, Q. Rev. Biophys., 35 (2002) 369-430.
213
[265] B.H. Stuart, Infrared Spectroscopy: Fundamentals and Applications, John Wiley & Sons Ltd, Chichester, West Sussex, 2004.
[266] A. Vainrub, O. Pustovyy, V. Vodyanoy, Resoltuion of 90 nm (x/5) in an Optical Transmission Microscope with an Annular Condenser, Optics Letters, 31 (2006) 2855-2857.
[267] D.C. Neckers, Rose bengal, J. Photochem. Photobiol. A: Chem., 47 (1989) 1-29.
[268] D. Jassby, J.F. Budarz, M.R. Wiesner, Impact of Aggregate Size and Structure on the Photocatalytic Properties of TiO2 and ZnO Nanoparticles, Environ. Sci. Technol., (2012) DOI: 10.1021/es202009h.
[269] T. Ashikaga, A. Motoyama, H. Ichikawa, H. Itagaki, Y. Sato, Low-molecular Peptide Model of Protein Denaturation by UVA Irradiation and the Effect of Antioxidants, Alter. Anim. Tes. and Exper., 7 (2000) 30-36.
[270] J.P. McCormick, T. Thomason, Near-Ultraviolet Photooxidation of Tryptophan. Proof of Formation of Superoxide Ion, J. Am. Chem. Soc., 100 (1978) 312-313.
[271] E.W. Blanch, L. Hecht, C.D. Syme, V. Volpetti, G.P. Lomonossoff, K. Nielsen, L.D. Barron, Molecular Structure of Viruses from Raman Optical Activity, J. Gen. Virol., 83 (2002) 2593-2600.
[272] D.A. Kuzmanovic, I. Elashvili, C. Wick, C. O'Connell, S. Krueger, The MS2 Coat Protein Shell is Likely Assembled Under Tension: A Novel Role for the MS2 Bacteriophage A Protein as Revealed by Small-angle Neutron Scattering, J. Mol. Biol., 355 (2006) 1095-1111.
[273] M.J. Davies, Singlet oxygen-mediated damage to proteins and its consequences, Biochem. Biophys. Res. Commun., 305 (2003) 761-770.
[274] E. Shacter, Quantification and significance of protein oxidation in biological samples, Drug Metabol. Rev., 32 (2000) 307-326.
[275] J.C. Crittenden, R.R. Trussell, D.W. Hand, K.J. Howe, G. Tchobanoglous, Water Treatment: Principles and Design, 2 ed., John Wiley & Sons, Inc., Hoboken, NJ, 2005.
[276] J. Lee, Y. Mackeyev, M. Cho, D. Li, J.H. Kim, L.J. Wilson, P.J.J. Alvarez, Photochemical and Antimicrobial Properties of Novel C(60) Derivatives in Aqueous Systems, Environmental Science & Technology, 43 (2009) 6604-6610.
[277] T. Kohn, K.L. Nelson, Sunlight-mediated inactivation of MS2 coliphage via exogenous singlet oxygen produced by sensitizers in natural waters, Environ. Sci. Technol., 41 (2007) 192-197.
[278] D.M. Guldi, P.V. Kamat, Photophysical properties of pristine fullerenes, functionalized fullerenes, and fullerene-contianing donor-bridge acceptor systems, in: K.M. Kadish, R.S. Ruoff (Eds.) Fullerenes, Wiley-Interscience, New York, NY, 2000, pp. 225-281.
214
[279] A. Weir, P. Westerhoff, L. Fabricius, K. Hristovski, N. Von Goetz, Titanium dioxide nanoparticles in food and personal care products, Environ Sci Technol, 46 (2012) 2242-2250.
[280] O. Carp, C.L. Huisman, A. Reller, Photoinduced reactivity of titanium dioxide, Progress in solid state chemistry, 32 (2004) 33-177.
[281] S. Malato, P. Fernández-Ibáñez, M. Maldonado, J. Blanco, W. Gernjak, Decontamination and disinfection of water by solar photocatalysis: recent overview and trends, Catalysis Today, 147 (2009) 1-59.
[282] H. Choi, S.R. Al-Abed, D.D. Dionysiou, E. Stathatos, P. Lianos, TiO 2-based advanced oxidation nanotechnologies for water purification and reuse, Sustainability science and engineering, 2 (2010) 229-254.
[283] M. Lu, P. Pichat, Photocatalysis and water purification: from fundamentals to recent applications, John Wiley & Sons, 2013.
[284] F. Loosli, P. Le Coustumer, S. Stoll, TiO 2 nanoparticles aggregation and disaggregation in presence of alginate and Suwannee River humic acids. pH and concentration effects on nanoparticle stability, Water Res, 47 (2013) 6052-6063.
[285] M. Abdullah, G.K. Low, R.W. Matthews, Effects of common inorganic anions on rates of photocatalytic oxidation of organic carbon over illuminated titanium dioxide, J Phys Chem-Us, 94 (1990) 6820-6825.
[286] H. Chen, O. Zahraa, M. Bouchy, Inhibition of the adsorption and photocatalytic degradation of an organic contaminant in an aqueous suspension of TiO 2 by inorganic ions, Journal of Photochemistry and Photobiology A: Chemistry, 108 (1997) 37-44.
[287] R.A. Burns, J.C. Crittenden, D.W. Hand, V.H. Selzer, L.L. Sutter, S.R. Salman, Effect of inorganic ions in heterogeneous photocatalysis of TCE, Journal of Environmental Engineering, 125 (1999) 77-85.
[288] M. Mehrvar, W.A. Anderson, M. Moo-Young, Photocatalytic degradation of aqueous organic solvents in the presence of hydroxyl radical scavengers, International Journal of photoenergy, 3 (2001) 187-191.
[289] K. Wang, J. Zhang, L. Lou, S. Yang, Y. Chen, UV or visible light induced photodegradation of AO7 on TiO 2 particles: the influence of inorganic anions, Journal of Photochemistry and Photobiology A: Chemistry, 165 (2004) 201-207.
[290] M.H. Habibi, A. Hassanzadeh, S. Mahdavi, The effect of operational parameters on the photocatalytic degradation of three textile azo dyes in aqueous TiO 2 suspensions, Journal of Photochemistry and Photobiology A: Chemistry, 172 (2005) 89-96.
215
[291] C. Guillard, E. Puzenat, H. Lachheb, A. Houas, J.-M. Herrmann, Why inorganic salts decrease the TiO2 photocatalytic efficiency, International Journal of Photoenergy, 7 (2005) 1-9.
[292] A. Lair, C. Ferronato, J.-M. Chovelon, J.-M. Herrmann, Naphthalene degradation in water by heterogeneous photocatalysis: an investigation of the influence of inorganic anions, Journal of Photochemistry and Photobiology A: Chemistry, 193 (2008) 193-203.
[293] M. Kerzhentsev, C. Guillard, J.-M. Herrmann, P. Pichat, Photocatalytic pollutant removal in water at room temperature: case study of the total degradation of the insecticide fenitrothion (phosphorothioic acid O, O-dimethyl-O-(3-methyl-4-nitro-phenyl) ester), Catalysis Today, 27 (1996) 215-220.
[294] G.E. Brown, V.E. Henrich, W.H. Casey, D.L. Clark, C. Eggleston, A. Felmy, D.W. Goodman, M. Grätzel, G. Maciel, M.I. McCarthy, Metal oxide surfaces and their interactions with aqueous solutions and microbial organisms, Chem Rev, 99 (1999) 77-174.
[295] M.A. Anderson, D.T. Malotky, The adsorption of protolyzable anions on hydrous oxides at the isoelectric pH, J Colloid Interf Sci, 72 (1979) 413-427.
[296] S.B. Johnson, G.E. Brown, T.W. Healy, P.J. Scales, Adsorption of organic matter at mineral/water interfaces. 6. Effect of inner-sphere versus outer-sphere adsorption on colloidal stability, Langmuir, 21 (2005) 6356-6365.
[297] U. Diebold, The surface science of titanium dioxide, Surf Sci Rep, 48 (2003) 53-229.
[298] A. Fernandez-Nieves, F. De Las Nieves, The role of ζ potential in the colloidal stability of different TiO 2/electrolyte solution interfaces, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 148 (1999) 231-243.
[299] R.A. French, A.R. Jacobson, B. Kim, S.L. Isley, R.L. Penn, P.C. Baveye, Influence of ionic strength, pH, and cation valence on aggregation kinetics of titanium dioxide nanoparticles, Environ Sci Technol, 43 (2009) 1354-1359.
[300] A.A. Keller, H. Wang, D. Zhou, H.S. Lenihan, G. Cherr, B.J. Cardinale, R. Miller, Z. Ji, Stability and aggregation of metal oxide nanoparticles in natural aqueous matrices, Environ Sci Technol, 44 (2010) 1962-1967.
[301] A.R. Petosa, D.P. Jaisi, I.R. Quevedo, M. Elimelech, N. Tufenkji, Aggregation and deposition of engineered nanomaterials in aquatic environments: role of physicochemical interactions, Environ Sci Technol, 44 (2010) 6532-6549.
[302] Y.H. Shih, W.S. Liu, Y.F. Su, Aggregation of stabilized TiO2 nanoparticle suspensions in the presence of inorganic ions, Environ Toxicol Chem, 31 (2012) 1693-1698.
216
[303] N. Solovitch, J.r.m. Labille, J.r.m. Rose, P. Chaurand, D. Borschneck, M.R. Wiesner, J.-Y. Bottero, Concurrent aggregation and deposition of TiO2 nanoparticles in a sandy porous media, Environ Sci Technol, 44 (2010) 4897-4902.
[304] M. Kosmulski, The significance of the difference in the point of zero charge between rutile and anatase, Adv Colloid Interfac, 99 (2002) 255-264.
[305] P. Velikovská, P. Mikulášek, The influence of Cl−, SO 4 2− and PO 4 3− ions on the ζ-potential and microfiltration of titanium dioxide dispersions, Separation and Purification Technology, 58 (2007) 295-298.
[306] B.P. Nelson, R. Candal, R.M. Corn, M.A. Anderson, Control of surface and ζ potentials on nanoporous TiO2 films by potential-determining and specifically adsorbed ions, Langmuir, 16 (2000) 6094-6101.
[307] V. Kazarinov, V. Andreev, A. Mayorov, Investigation of the adsorption properties of the TiO2 electrode by the radioactive tracer method, Journal of Electroanalytical Chemistry and Interfacial Electrochemistry, 130 (1981) 277-285.
[308] W. Stumm, The inner-sphere surface complex: A key to understanding surface reactivity, Advances in chemistry series, 244 (1995) 1-1.
[309] L.-F. Liao, C.-F. Lien, D.-L. Shieh, M.-T. Chen, J.-L. Lin, FTIR study of adsorption and photoassisted oxygen isotopic exchange of carbon monoxide, carbon dioxide, carbonate, and formate on TiO2, The Journal of Physical Chemistry B, 106 (2002) 11240-11245.
[310] R. Luschtinetz, J. Frenzel, T. Milek, G. Seifert, Adsorption of phosphonic acid at the TiO2 anatase (101) and rutile (110) surfaces, The Journal of Physical Chemistry C, 113 (2009) 5730-5740.
[311] K.D. Kwon, J.D. Kubicki, Molecular orbital theory study on surface complex structures of phosphates to iron hydroxides: Calculation of vibrational frequencies and adsorption energies, Langmuir, 20 (2004) 9249-9254.
[312] S.J. Hug, D. Bahnemann, Infrared spectra of oxalate, malonate and succinate adsorbed on the aqueous surface of rutile, anatase and lepidocrocite measured with in situ ATR-FTIR, Journal of electron spectroscopy and related phenomena, 150 (2006) 208-219.
[313] J.-M. Herrmann, P. Pichat, Heterogeneous photocatalysis. Oxidation of halide ions by oxygen in ultraviolet irradiated aqueous suspension of titanium dioxide, Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases, 76 (1980) 1138.
[314] A. Rincon, Effect of pH, inorganic ions, organic matter and H2O2 on E. coli K12 photocatalytic inactivation by TiO2Implications in solar water disinfection, Applied Catalysis B: Environmental, 51 (2004) 283-302.
217
[315] Y. Katsumura, P. Jiang, R. Nagaishi, T. Oishi, K. Ishigure, Y. Yoshida, Pulse radiolysis study of aqueous nitric acid solutions: formation mechanism, yield, and reactivity of NO3 radical, The Journal of Physical Chemistry, 95 (1991) 4435-4439.
[316] J. Kochany, E. Lipczynska-Kochany, Application of the EPR spin-trapping technique for the investigation of the reactions of carbonate, bicarbonate, and phosphate anions with hydroxyl radicals generated by the photolysis of H 2 O 2, Chemosphere, 25 (1992) 1769-1782.
[317] J. Holcman, T. Løgager, K. Sehested, Reactions of peroxodisulphate and SO 4-with selected radicals in aqueous solution, Laboratory Studies on Atmospheric Chemistry, (1992).
[318] L.E. Gomez, Use and benefits of chlorpyrifos in US agriculture, Dow AgroSciences LLC, (2009).
[319] Z. Chishti, S. Hussain, K.R. Arshad, A. Khalid, M. Arshad, Microbial degradation of chlorpyrifos in liquid media and soil, Journal of environmental management, 114 (2013) 372-380.
[320] S. Armenta, G. Quintás, S. Garrigues, M. de la Guardia, A validated and fast procedure for FTIR determination of Cypermethrin and Chlorpyrifos, Talanta, 67 (2005) 634-639.
[321] S.Y. Gebremariam, M.W. Beutel, D.R. Yonge, M. Flury, J.B. Harsh, Adsorption and Desorption of Chlorpyrifos to Soils and Sediments, Rev Environ Contam T, 215 (2012) 123-175.
[322] S. Baskaran, R.S. Kookana, R. Naidu, Contrasting behaviour of chlorpyrifos and its primary metabolite, TCP (3,5,6-trichloro-2-pyridinol), with depth in soil profiles, Aust J Soil Res, 41 (2003) 749-760.
[323] R.J. Gilliom, J.E. Barbash, C.G. Crawford, P.A. Hamilton, J.D. Martin, N. Nakagaki, L.H. Nowell, J.C. Scott, P.E. Stackelberg, G.P. Thelin, Pesticides in the nation's streams and ground water, 1992-2001, Geological Survey (US), 2006.
[324] E. Estevez, M. del Carmen Cabrera, J.R. Fernández-Vera, A. Molina-Díaz, J. Robles-Molina, M. del Pino Palacios-Díaz, Monitoring priority substances, other organic contaminants and heavy metals in a volcanic aquifer from different sources and hydrological processes, Sci Total Environ, 551 (2016) 186-196.
[325] M.P. Ensminger, R. Budd, K.C. Kelley, K.S. Goh, Pesticide occurrence and aquatic benchmark exceedances in urban surface waters and sediments in three urban areas of California, USA, 2008–2011, Environmental monitoring and assessment, 185 (2013) 3697-3710.
[326] K.D. Racke, D.D. Fontaine, R.N. Yoder, J.R. Miller, Chlorpyrifos degradation in soil at termiticidal application rates, Pestic Sci, 42 (1994) 43-51.
[327] K.R. Solomon, W.M. Williams, D. Mackay, J. Purdy, J.M. Giddings, J.P. Giesy, Properties and uses of chlorpyrifos in the United States, Reviews of environmental contamination and toxicology, 231 (2014) 13-34.
218
[328] A.W. Jarvinen, B.R. Nordling, M.E. Henry, Chronic toxicity of Dursban (chlorpyrifos) to the fathead minnow (Pimephales promelas) and the resultant acetylcholinesterase inhibition, Ecotoxicology and environmental safety, 7 (1983) 423-434.
[329] J.P. Giesy, K.R. Solomon, J.R. Coats, K.R. Dixon, J.M. Giddings, E.E. Kenaga, Chlorpyrifos: ecological risk assessment in North American aquatic environments, Springer, 1999.
[330] D.L. Eaton, R.B. Daroff, H. Autrup, J. Bridges, P. Buffler, L.G. Costa, J. Coyle, G. McKhann, W.C. Mobley, L. Nadel, D. Neubert, R. Schulte-Hermann, P.S. Spencer, Review of the toxicology of chlorpyrifos with an emphasis on human exposure and neurodevelopment, Critical reviews in toxicology, 38 Suppl 2 (2008) 1-125.
[331] M.E. DeLorenzo, G.I. Scott, P.E. Ross, Toxicity of pesticides to aquatic microorganisms: a review, Environ Toxicol Chem, 20 (2001) 84-98.
[332] L.G. Devi, B.N. Murthy, S.G. Kumar, Photocatalytic activity of V5+, Mo6+ and Th4+ doped polycrystalline TiO2 for the degradation of chlorpyrifos under UV/solar light, Journal of Molecular Catalysis A: Chemical, 308 (2009) 174-181.
[333] J. Risher, H.A. Navarro, Toxicological profile for chlorpyrifos, US Department of Health and Human Services, Altanta, (1997).
[334] D. Qiao, F.J. Seidler, T.A. Slotkin, Developmental neurotoxicity of chlorpyrifos modeled in vitro: Comparative effects of metabolites and other cholinesterase inhibitors on DNA synthesis in PC12 and C6 cells, Environ Health Persp, 109 (2001) 909-913.
[335] S. Baskaran, R.S. Kookana, R. Naidu, Degradation of bifenthrin, chlorpyrifos and imidacloprid in soil and bedding materials at termiticidal application rates, Pestic Sci, 55 (1999) 1222-1228.
[336] E.M. John, J.M. Shaike, Chlorpyrifos: pollution and remediation, Environmental Chemistry Letters, 13 (2015) 269-291.
[337] J.M. Smolen, A.T. Stone, Metal (hydr)oxide surface catalyzed hydrolysis of chlorpyrifos-methyl, chlorpyrifos-methyl oxon, and paraoxon, Soil Sci Soc Am J, 62 (1998) 636-643.
[338] M.T. Rose, A.N. Crossan, I.R. Kennedy, Dissipation of cotton pesticides from runoff water in glasshouse columns, Water, air, and soil pollution, 182 (2007) 207-218.
[339] D.S. Bhatkhande, V.G. Pangarkar, A.A.C.M. Beenackers, Photocatalytic degradation for environmental applications - a review, Journal of Chemical Technology & Biotechnology, 77 (2002) 102-116.
[340] U.I. Gaya, A.H. Abdullah, Heterogeneous photocatalytic degradation of organic contaminants over titanium dioxide: A review of fundamentals, progress and problems, Journal of Photochemistry and Photobiology C: Photochemistry Reviews, 9 (2008) 1-12.
219
[341] C. McCullagh, N. Skillen, M. Adams, P.K.J. Robertson, Photocatalytic reactors for environmental remediation: a review, Journal of Chemical Technology & Biotechnology, 86 (2011) 1002-1017.
[342] M.M. Khin, A.S. Nair, V.J. Babu, R. Murugan, S. Ramakrishna, A review on nanomaterials for environmental remediation, Energy & Environmental Science, 5 (2012) 8075.
[343] S. Kwon, M. Fan, A.T. Cooper, H. Yang, Photocatalytic Applications of Micro- and Nano-TiO2in Environmental Engineering, Critical Reviews in Environmental Science and Technology, 38 (2008) 197-226.
[344] A. Dârjan, C. Drăghici, D. Perniu, A. Duţă, Degradation of Pesticides by TiO2 Photocatalysis, (2013) 155-163.
[345] T. Ochiai, A. Fujishima, Photoelectrochemical properties of TiO2 photocatalyst and its applications for environmental purification, Journal of Photochemistry and Photobiology C: Photochemistry Reviews, 13 (2012) 247-262.
[346] A.C. Affam, M. Chaudhuri, Degradation of pesticides chlorpyrifos, cypermethrin and chlorothalonil in aqueous solution by TiO 2 photocatalysis, Journal of environmental management, 130 (2013) 160-165.
[347] A. Fadaei, M. Kargar, Photocatalytic degradation of chlorpyrifos in water using titanium dioxide and zinc oxide, Fresenius Environmental Bulletin, 22 (2013) 2442-2447.
[348] A.S. Nair, T. Pradeep, Extraction of chlorpyrifos and malathion from water by metal nanoparticles, J Nanosci Nanotechno, 7 (2007) 1871-1877.
[349] A. Torrents, A.T. Stone, Oxide surface-catalyzed hydrolysis of carboxylate esters and phosphorothioate esters, Soil Sci Soc Am J, 58 (1994) 738-745.
[350] L. Clausen, I. Fabricius, Atrazine, isoproturon, mecoprop, 2, 4-D, and bentazone adsorption onto iron oxides, J Environ Qual, 30 (2001) 858-869.
[351] T. Hofmann, F. von der Kammer, Estimating the relevance of engineered carbonaceous nanoparticle facilitated transport of hydrophobic organic contaminants in porous media, Environmental pollution, 157 (2009) 1117-1126.
[352] M.N. Moore, Do nanoparticles present ecotoxicological risks for the health of the aquatic environment?, Environ Int, 32 (2006) 967-976.
[353] L. Wang, Y. Huang, A.T. Kan, M.B. Tomson, W. Chen, Enhanced transport of 2,2',5,5'-polychlorinated biphenyl by natural organic matter (NOM) and surfactant-modified fullerene nanoparticles (nC60), Environ Sci Technol, 46 (2012) 5422-5429.
220
[354] N. Pino, G. Peñuela, Simultaneous degradation of the pesticides methyl parathion and chlorpyrifos by an isolated bacterial consortium from a contaminated site, International Biodeterioration & Biodegradation, 65 (2011) 827-831.
[355] B.K. Singh, A. Walker, J.A. Morgan, D.J. Wright, Biodegradation of chlorpyrifos by enterobacter strain B-14 and its use in bioremediation of contaminated soils, Applied and environmental microbiology, 70 (2004) 4855-4863.
[356] B.K. Singh, A. Walker, J.A.W. Morgan, D.J. Wright, Effects of Soil pH on the Biodegradation of Chlorpyrifos and Isolation of a Chlorpyrifos-Degrading Bacterium, Applied and environmental microbiology, 69 (2003) 5198-5206.
[357] J.M. Bifulco, J. Shirey, G. Bissonnette, Detection of Acinetobacter spp. in rural drinking water supplies, Applied and environmental microbiology, 55 (1989) 2214-2219.
[358] S. Anwar, F. Liaquat, Q.M. Khan, Z.M. Khalid, S. Iqbal, Biodegradation of chlorpyrifos and its hydrolysis product 3,5,6-trichloro-2-pyridinol by Bacillus pumilus strain C2A1, J Hazard Mater, 168 (2009) 400-405.
[359] F. Ahmad, S. Iqbal, S. Anwar, M. Afzal, E. Islam, T. Mustafa, Q.M. Khan, Enhanced remediation of chlorpyrifos from soil using ryegrass (Lollium multiflorum) and chlorpyrifos-degrading bacterium Bacillus pumilus C2A1, J Hazard Mater, 237 (2012) 110-115.
[360] Y. Ge, J.P. Schimel, P.A. Holden, Evidence for negative effects of TiO2 and ZnO nanoparticles on soil bacterial communities, Environ Sci Technol, 45 (2011) 1659-1664.
[361] G. Imfeld, S. Vuilleumier, Measuring the effects of pesticides on bacterial communities in soil: A critical review, European Journal of Soil Biology, 49 (2012) 22-30.
[362] P. Palma, V.L. Palma, R.M. Fernandes, A.M. Soares, I.R. Barbosa, Acute toxicity of atrazine, endosulfan sulphate and chlorpyrifos to Vibrio fischeri, Thamnocephalus platyurus and Daphnia magna, relative to their concentrations in surface waters from the Alentejo region of Portugal, Bull Environ Contam Toxicol, 81 (2008) 485-489.
[363] T.J. Battin, F.V. Kammer, A. Weilhartner, S. Ottofuelling, T. Hofmann, Nanostructured TiO2: transport behavior and effects on aquatic microbial communities under environmental conditions, Environ Sci Technol, 43 (2009) 8098-8104.
[364] L.K. Adams, D.Y. Lyon, P.J. Alvarez, Comparative eco-toxicity of nanoscale TiO 2, SiO 2, and ZnO water suspensions, Water Res, 40 (2006) 3527-3532.
[365] M. Biguzzi, G. Shama, Effect of titanium dioxide concentration on the survival of Pseudomonas stutzeri during irradiation with near ultraviolet light, Letters in Applied Microbiology, 19 (1994) 458-460.
221
[366] D.Y. Lyon, L.K. Adams, J.C. Falkner, P.J.J. Alvarez, Antibacterial activity of fullerene water suspensions: Effects of preparation method and particle size, Environ Sci Technol, 40 (2006) 4360-4366.
[367] V.G.G. Kanmoni, S. Daniel, G.A.G. Raj, Photocatalytic degradation of chlorpyrifos in aqueous suspensions using nanocrystals of ZnO and TiO2, Reaction Kinetics, Mechanisms and Catalysis, 106 (2012) 325-339.
[368] K. Sivagami, R.R. Krishna, T. Swaminathan, Photo Catalytic Degradation of Chlorpyriphos in an Annular Slurry Reactor, Journal of Water Sustainability, 3 (2013) 143-151.
[369] A. Verma, D. Dixit, Photocatalytic degradability of insecticide Chlorpyrifos over UV irradiated Titanium dioxide in aqueous phase, International Journal of Environmental Sciences, 3 (2012) 743-755.
[370] M.J. Redondo, M.J. Ruiz, G. Font, R. Boluda, Dissipation and distribution of atrazine, simazine, chlorpyrifos, and tetradifon residues in citrus orchard soil, Arch Environ Con Tox, 32 (1997) 346-352.
[371] S.E. Duirk, T.W. Collette, Degradation of chlorpyrifos in aqueous chlorine solutions: Pathways, kinetics, and modeling, Environ Sci Technol, 40 (2006) 546-551.
[372] K.L. Armbrust, Pesticide hydroxyl radical rate constants: measurements and estimates of their importance in aquatic environments, Environ Toxicol Chem, 19 (2000) 2175-2180.
[373] F. Wilkinson, W.P. Helman, A.B. Ross, Rate constants for the decay and reactions of the lowest electronically excited state of molecular oxygen in solution. An expanded and revised compliation, Journal of Physical and Chemical Reference Data, 34 (1995) 663-1021.
[374] A.A. Green, M. Berman, P. Switzer, M.D. Craig, A Transformation for ordering multispectral data in terms of image quality with implications for noise removal, IEEE Trans. Geosci. Rem. Sen., 26 (1988) 65-74.
[375] U.K. Laemmli, Cleavage of structural proteins during the assembly of the head of bacteriophage T4, Nature, 227 (1970) 680-685.
[376] I. Grc ić, G. Li Puma, Photocatalytic degradation of water contaminants in multiple photoreactors and evaluation of reaction kinetic constants independent of photon absorption, irradiance, reactor geometry, and hydrodynamics, Environ Sci Technol, 47 (2013) 13702-13711.
222
Biography Jeffrey Michael Farner Budarz was born on September 14, 1981 in Lafayette,
Indiana. He received his BS in Chemistry from Purdue University in December of 2004
following interludes at Barrett, the Honors College at Arizona State University and
Université Louis Pasteur in Strasbourg, France. After obtaining his undergraduate
degree, he worked for Iowa’s State Hygienic Lab, as an analytical chemist, performing
water quality analyses for compliance with the EPA’s SDWA and FIFRA. In August of
2008, he began his graduate work at Duke University in the Department of Civil and
Environmental Engineering. He has authored or co-authored the following articles:
Experimental Measurement and Modeling of Reactive Species Generation in TiO2
Nanoparticle Photocatalysis. Chem Eng J, 2015; Formation and Persistence of Silver
Nanoparticles in Visible Light Illuminated Aquatic Systems and their Effect on Bacterial
Lysis. EES, 2014; Bacteriophage Inactivation by UV-A Illuminated Fullerenes: Role of
Nanoparticle-Virus Association and Biological Targets. ES&T, 2012; Impact of Aggregate
Size and Structure on the Photocatalytic Properties of TiO2 and ZnO Nanoparticles.
ES&T, 2012; Comparison of the Photosensitivity and Bacterial Toxicity of Spherical and
Tubular Fullerenes of Variable Aggregate Size. J Nano Res, 2011; Heterogeneities in
Fullerene Nanoparticle Aggregates Affecting Reactivity, Bioactivity, and Transport. ACS
Nano, 2010. During his graduate work, Jeff received the Jeffrey B. Taub award from the
Department of Civil and Environmental Engineering in 2011 and 2012.